Pricing activity is a running log of the material changes we track across the Blueprint — price moves, packaging and plan changes, product launches, funding, and deprecations. Newest first, each linked to the company it affects.
As of , UsagePricing has tracked 506 material pricing changes across 230 AI companies — 83 launches, 121 price moves, 228 packaging changes, 53 funding events, 13 deprecations, 1 hiring, 5 stacks.
Updated
Anthropic cancelled Sonnet 5's scheduled Sept 1 price hike — the $2/$10 rate is now permanent, breaking the $3/$15 Sonnet anchor held since Claude 3.
DetailsApollo lowered its per-minute Dialer credit cost in three regions — Middle East 35 to 28 credits/min, Africa 45 to 36, Central & South America 15 to 14 — while US, North America, Europe, and APAC rates held steady, confirmed via Apollo's live pricing API.
DetailsBaseten grew its Model APIs catalog from 12 to 14 models, adding GLM-5.3-Flash ($0.15/$0.03/$0.50 per 1M tokens) and DeepSeek V4 Pro 0813 ($1.32/$0.132/$3.96); no existing SKU's rate changed.
DetailsBright Data restructured Browser API from a fixed $8-to-$5/GB, four-tier committed ladder to a Free / PAYG ($8/GB) / Scale ($499/mo, slider-adjustable) / Enterprise model, mirroring its July 2026 API restructure.
DetailsBright Data's Managed Services (formerly Managed Data Acquisition) now shows a two-tier $1,000/mo or $2,500/mo structure plus a $500 setup fee and $4/1k-request usage, replacing the single $1,500/mo figure on record.
DetailsCartesia's Monthly/Yearly (Save 20%) billing toggle is gone from cartesia.ai/pricing again, just two days after being reinstated — every self-serve tier now shows one flat monthly rate: Pro $5, Startup $49, Scale $299.
DetailsCoreWeave added a spot-only Standard Memory configuration of its RTX PRO 6000 Blackwell SE GPU shape at $9.56/hr, gating its on-demand rate to Contact sales while the existing High Memory configuration held at $20.00/hr.
DetailsDeepInfra's /pricing page now lists only claude-haiku-4-5 ($1.00/$5.00 per 1M); six other Claude models, including the $10.00/$50.00 claude-fable-5 row, are gone, down from seven models shown as of 2026-08-26.
DetailsDeepInfra added a 35%-off promo tag to GLM-5.2, cutting its post-2026-08-04-cut rate of $0.75/$2.40 ($0.14 cached) to $0.488/$1.56 ($0.091 cached) per 1M tokens on the /models and /deepstart catalogs.
DetailsFrase tripled-to-sextupled Content Guard page caps per tier and let every plan, including Starter, auto-publish any content type -- with no price change.
DetailsGitLab's AI credit add-on (GitLab Duo Agent Platform, $1/credit) is now available with any plan, including Free, not just Premium/Ultimate as before, and GitLab launched a new Secrets Manager add-on for Premium/Ultimate — core seat and credit list prices stayed flat.
DetailsHeidi Health cut its plan ladder from six tiers to four, dropping Evidence Plus/Team and moving Practice from a fixed $180/user/month to a sales-quoted price.
DetailsHeptabase now lets any paid tier enable on-demand usage with a user-set monthly spending limit, so extra AI credit consumption is billed pay-as-you-go instead of forcing a full tier upgrade.
DetailsHiggsfield removed the $3 'Special' one-time credit pack that had briefly undercut its $9 Basic plan, along with the credits-per-dollar comparison strip it added the same week — subscription tier prices are unchanged.
DetailsHippocratic AI's site now leads with an 'Orchestrators' framing and dozens of newly named Skills across Provider, Payor and Life Sciences; its per-agent-hour meter (~$9/hour) is unchanged.
DetailsLanceDB dropped the self-serve 'Cloud' product name from its docs and marketing; the only paid tier is now Enterprise, sold via annual commit as Managed or BYOC deployment.
DetailsLinkup's billing docs add a second Fetch pricing dimension - a 'pro' mode alongside the existing JS-rendering toggle - plus a flat $0.001 structured-output surcharge, pushing the top Fetch rate from $0.005 to $0.011.
DetailsLiveKit Inference gained two speech-to-text SKUs, Google Gemini 3.5 Transcribe Live and Speechmatics Linden-1 (both $0.0095/min and $0.0050/min flat), while the catalog's xAI vendor listing was relabeled SpaceXAI; all LiveKit Cloud plan prices held unchanged.
DetailsMaven AGI shipped two newly named platform capabilities, Charters and Intelligent Fields, renamed its knowledge layer to Graph of Record, and broadened its marketed use cases from three customer-support categories to nine spanning the full customer journey.
DetailsA new Business tab on Kimi's membership page reveals a self-serve $600/year/seat price (min. 2 seats) with a working Subscribe checkout, three weeks after the tier was pulled sales-only.
DetailsPixee promoted Foresight, its PRD/design-stage security review agent, from Early Access to Live, completing the reactive-plus-proactive pairing on its Context Graph platform with no change to its quote-only pricing.
DetailsRecraft's pricing page and docs both dropped the specific '30 credits per day' figure for the Free tier, now describing it only as 'limited free credits' with no numeric daily allowance stated.
DetailsReplit lowered Core to $20/mo monthly ($18 annual, down from $25/$20) and Pro to $90/mo annual (down from $95), replacing its prior 20%/5% annual discounts with a uniform 10% off across both paid tiers.
DetailsSarvam's marketing pricing page was rebuilt around a 2-card Starter/Enterprise hero plus a 5-tier per-unit rate table; its month-old 7x LLM price gap vs the developer docs is now resolved in the docs' favor, while Translate and Dubbing show new, higher, tiered rates that disagree with docs instead.
DetailsTavus's dedicated tavus.io/pals landing page now returns a branded 404 on both the bare and www domains, after carrying no list prices as recently as July 29 — the PAL Maker tool and PAL-related FAQ entries stay live on the developer pricing page.
DetailsUniphore introduced Marketing AI on August 17, 2026 — a Business AI Cloud product that builds per-customer predictive models to simulate marketing outcomes before spend. The launch adds a new front-office product line but ships with no published pricing, sold under the same sales-led, quote-only motion as the rest of the platform.
DetailsUnstructured cut its Pay-As-You-Go rate from $0.03 to $0.015 per page and reverted the 15,000-page monthly free allowance to a one-time 10,000-page signup allowance, removing the $3,000/mo bill cap it had added just six weeks earlier.
DetailsVellum renamed the Custom plan's optional recurring credit top-up from 'credit bundle' to 'usage bundle' and cut its menu from five flexible amounts to three fixed tiers matching the Mighty/Super/Ultra presets, while grandfathering existing subscribers' old bundles.
DetailsWeaviate launched Engram, managed AI-agent memory billed per pipeline run, on a Free/Starter($45)/Team($360)/Enterprise ladder, separate from database pricing.
DetailsUnbabel's self-serve AI-translation product Widn.ai is scheduled to be decommissioned on 2026-08-27 and replaced by TransPerfect's GlobalLink Now. Widn.ai was Unbabel's only public price ladder (Free / $19 / $90 / Custom); its retirement returns the company to a fully sales-gated posture under the GlobalLink brand following the August 2025 TransPerfect acquisition.
DetailsGroq's docs added a price for Qwen/Qwen3.8-27B in the Preview Models table — $0.80 input / $4.00 output per 1M tokens at 450 T/SEC — a model that previously had rate limits but no listed price.
DetailsHubSpot's AI Answer Engine Optimization tool now carries a published standalone price ($50/mo, $45/mo annual) after appearing only as an unpriced Beta feature bundled into Agent Hub since its 2026-07-30 launch.
DetailsOne day after syncing at $19/seat, intercom.com/pricing's annual toggle reverts Essential to the clean $29/seat/mo list rate — but the site's own calculator and fin.ai/pricing still show the discounted $19/seat/mo, the first split between two same-domain surfaces in this saga.
DetailsMetronome's Startup plan no longer includes a free US$100,000 billing-volume / 10-million-event allotment — its 0.8% and $0.04-per-1,000-event rates now apply to all usage from the first dollar and first event, not just overages.
DetailsWix's entry-level Light Premium plan rose 12% from $17/mo to $19/mo (yearly billing, paid upfront); Core ($29), Business ($39, RECOMMENDED) and Business Elite ($159) held steady.
DetailsApify's Free plan now caps Actor runs at 5 concurrent, down from 25 in the prior capture, while every other plan's concurrency ceiling (32/128/256 on Starter/Scale/Business) is unchanged.
DetailsAssemblyAI now itemizes its $4.50/hr Voice Agent bundle (STT, LLM, TTS all 'Included') and confirms via FAQ there are no per-layer add-on fees on top of that rate.
DetailsAugment added Gemini 3.7 Flash and Grok 4.6 (its first xAI model) to its per-token rate table; existing model rates and the $100/mo Business plan are unchanged.
DetailsWordPress.com's billing selector grew from 3 periods to 4 with a new 'Pay every 3 years' option (Personal drops to $2.75/mo), and forcing USD pricing on cloud.jetpack.com revealed real prices for Security, Growth, and Complete bundles that previously rendered as placeholders.
DetailsCartesia reinstated the Monthly/Yearly (Save 20%) toggle on its pricing page a month after removing it, defaulting again to yearly rates: Pro $4/mo, Startup $39/mo, Scale $239/mo (vs. $5/$49/$299 monthly).
DetailsCerebras removed ZAI-GLM-4.7 ($2.25/$2.75 per 1M tokens) from its public Developer Tier Pricing rate card and docs Model Catalog on its previously disclosed Aug 17, 2026 deprecation date, leaving two public models.
DetailsCerebras unveiled CS-4, built on three Wafer Scale Engine 3 Turbo processors on a new modular Nexus rack, claiming up to 30x faster inference than GPUs and up to 2x faster than CS-3; pricing is undisclosed and sold via enterprise contract.
DetailsAnthropic made Claude Sonnet 5's introductory $2/$10 per-MTok API price permanent, canceling the previously scheduled Sept 1, 2026 increase to $3/$15.
DetailsCursor's Auto Cost, Balance, and Intelligence modes now all bill at the routed model's own list price; the old flat $1.25 in / $6 out blended rate survives only as 'Legacy Enterprise Auto,' sunsetting September 7, 2026.
DetailsCursor's Other Models pool got cheaper: Claude Sonnet 5 down ~33% (input $3 to $2, output $15 to $10 per 1M tokens), GPT-5.6 Sol down ~20-33% ($5/$30 to $4/$20), and Gemini 3.7 Flash replaced Gemini 3.6 Flash at roughly half the per-token rate.
DetailsDeepgram added a new text-to-speech model, Flux TTS, running free (up to 45 concurrent streaming connections globally, 5 in EU/AU) through 9/12/2026, then $0.0450/1k characters Pay-As-You-Go ($0.0405/1k Growth) from 9/13/2026.
DetailsDeepgram's Voice Agent API 'Custom - BYO LLM' tier's standing Pay-As-You-Go rate rose from $0.056/min to $0.065/min (+16%); the Growth rate for the same tier is unchanged at $0.059/min.
DetailsDeepInfra cut MiMo-V2.5-Pro pricing 61% via a launch promo: $1.00/$3.00 ($0.20 cached) list price down to $0.39/$1.17 ($0.078 cached) per 1M tokens, shown as a struck-through price on the /models and /deepstart catalogs.
DetailsElementor introduced a new Expert plan at $17/mo (25 sites, $204/yr) between the existing Advanced tier and the Elementor One bundle, revealed by an in-card toggle rather than a separate pricing page.
DetailsExa's pricing page now names Starter/Developer/Enterprise tiers, disclosing Search QPS and Agent-concurrency caps plus a new Exa Connect pricing section.
DetailsGroq's two Llama models lost their public per-token price and now show Contact Sales in the docs, dropping off the Free and Developer rate-limit tables entirely.
DetailsGumloop replaced its $37-$1,840 credit-volume slider with one flat $37/mo Pro plan, added a new 8% agent orchestration fee (16% with BYOK), and re-capped overage after a March 2026 unlimited-by-default policy.
DetailsThe Essential new-customer offer that went dark 2026-08-14 is live again: $29 struck to $19/seat/mo annual (Save 35%) and $39 struck to $29/seat/mo monthly (Save 25%). For the first time this cycle, fin.ai/pricing's Intercom-helpdesk seat anchor also shows $19, in sync with the main site.
DetailsKrisp raised Call Center AI CC Core 50% (from $10 to $15/agent/mo) and bundled Voice Translation, Speech Analytics, and Voice Security into CC Advanced.
DetailsLinear's Coding Sessions AI-credit billing shifted from opaque task-tier estimates ($0.50-$5+) to itemized pricing: model tokens at provider-published rates with no markup plus $0.25 per 20-minute sandbox block.
DetailsLiveKit's per-minute inference catalog lost xAI's Grok 4.1 Fast and Grok 4.1 Fast Reasoning LLM SKUs and replaced Rime's Arcana voice with a new Gradium TTS provider; all LiveKit Cloud plan prices (Build free, Ship $50, Scale $500, Enterprise custom) held unchanged.
DetailsLlamaIndex's credit-pricing docs added a new Turbo (preview) all-in extraction rate of 35 credits/page with no separate parse step, plus a documented spreadsheet-mode rule billing workbooks at the Agentic Plus extract rate.
DetailsLlamaIndex's Pro plan ($500/mo, 400K credits) now advertises a one-time limited-time bonus of 800K credits ($1,000 value) for new signups on business email domains, forfeited immediately on downgrade.
DetailsMiniMax added a dedicated Token Plan for Teams surface (per-seat licensing plus a shared Credits pool) and closed its paid Music and Lyrics Generation APIs to new signups, grandfathering existing customers.
DetailsMiniMax's Token Plan subscriptions rose 10% on the international USD card — Plus $20→$22, Max $50→$55, and Ultra $120→$132 per month — with quotas and agent-concurrency limits unchanged.
DetailsMintlify now pools AI credits into one org-wide balance across all Pro deployments, replacing per-deployment tracking — an allocation change, not a price move.
DetailsMistral's API catalog grew: a third-party GLM 5.2 model, a quoted Enterprise APIs surcharge tier, a Priority inference speed, a free Leanstral coding agent, and a Data capture tool — while Vibe consumer plans held flat.
DetailsMistral repriced its content-moderation classifier from $0.1 per million input tokens to Free, relaunching it as Mistral Moderation 2 alongside a wider API catalog refresh.
DetailsDatabricks expanded Mosaic AI's Foundation Model APIs catalog with GLM-5.2, DeepSeek V4 Pro/Flash, Kimi K3/K2.7, and Qwen 3.5 122B, and added H100 GPU configurations to Model Serving — all still billed at the existing $0.07/DBU rate.
DetailsNetlify's AI Gateway repriced multiple OpenRouter-routed models by double digits (DeepSeek, Kimi, Qwen) and grew the catalog from ~161 to ~167 models, while its own plan prices stayed flat.
DetailsNovita re-added NVIDIA H100 SXM 80GB ($3.39/hr on-demand, $1.70/hr spot) and L40S 48GB ($0.55/hr) to its self-serve GPU-instance lineup, removing RTX 6000 Ada 48GB and the RTX 5090 'High frequency' variant in the same pass.
DetailsOpenAI repriced its GPT-5.6 API line on Aug 26, 2026: Luna fell 80% to $0.20/$1.20 per 1M tokens, Terra to $2.00/$12.00, and Sol to $4.00/$20.00 as promotional pricing through Nov 21, 2026.
DetailsRecraft added Recraft V4 Styles, V4 Styles Vector, V4 Styles Pro, and V4 Styles Pro Vector models that turn 1-10 reference images or SVGs into a reusable style, priced from $0.035 to $0.12 per image on the existing API-unit system.
DetailsRox restructured its three-tier pricing: the old $50/mo Core plan became a $100/mo Individual plan (still 10,000 actions per 5,000 at the old rate) and a new $255/mo Teams plan with unlimited seats and CRM sync now sits between it and Enterprise.
DetailsSemrush's pricing page now includes a permanent $0/mo Free plan (1 demo project, 10 reports/day, no credit card) alongside its four paid SEO + AI Search tiers, replacing trial-only free access.
DetailsStability AI added a $19/month Creator tier to Brand Studio and made its Platform API per-service credit pricing publicly visible without requiring sign-in.
Detailsturbopuffer now lists a separate +$128/mo Audit Log Streams add-on on its Scale and Enterprise tiers and raised the namespace pinning billing minimum from 64 GB to 128 GB, per its August 18, 2026 pricing changelog update.
DetailsVercel Sandbox expands from one region (iad1) to four (iad1, cle1, cdg1, sfo1), disclosing its first regional price gap: Active CPU and Provisioned Memory run about 38% higher in Paris and San Francisco than in Cleveland or Washington D.C.
DetailsVercel's pricing page drops the 'Vercel Connect' line item (5K requests/month included, $3 per 10K requests) and replaces it with two separately metered AI Gateway rows: Token Requests ($3.00 per 1,000 requests) and Triggers ($0.95 per 1,000 triggers).
DetailsApify started billing Unblocker, an anti-blocking proxy product, at $1.5/$1.5/$1.25/$1 per 1,000 requests across Free/Starter/Scale/Business, a new line item absent from the prior capture.
DetailsComposio's live pricing diverges from its July plan: no Business tier, Free jumps to 100K tool calls/mo, overage lands near $0.30/1K vs. the announced $4/1K.
DetailsFireworks repriced DeepSeek V4 Flash (0731): Standard input rose from $0.14 to $0.22 per 1M tokens (+57%), cached input fell from $0.028 to $0.007 (-75%), and output rose from $0.28 to $0.66 (+136%), with no forward-notice framing on the page.
DetailsGoogle launched Gemini 3.7 Flash and moved it and Gemini 3.6 Flash onto shared introductory pricing of $0.75/$3.75 per 1M tokens through Dec 31, 2026 — half of 3.6 Flash's prior $1.50/$7.50 rate, reverting Jan 1, 2027.
DetailsIdeogram halved Ideogram 4.0's in-app credit cost (2/4/6 to 1/2/3 credits per image) and its custom-model inference rate, added a 4.0 transparent-generation option, and folded API pricing into the main Plans & pricing page.
DetailsMem retired its flat $12/mo Pro plan and $99/mo Proactive tier for a 7-tier ladder (Free-$199/mo) capping messages, notes, tasks and Voice Mode on every plan.
DetailsMercor's advertised 'average contracted rate' fell from $114/hr to $109/hr on roles created rising to 401.9K, its lowest reading since the mid-July rebound. The Enterprise page's four offering-step demos were also visually refreshed; buyer pricing stays fully undisclosed.
DetailsNetlify's AI Gateway and Agent Runners now route to OpenRouter, adding roughly 161 open-weight and third-party models (DeepSeek, Qwen, Kimi, Llama, GLM, Grok, and more) at the same 180-credits-per-$1 metered rate as Anthropic, Gemini, and OpenAI.
DetailsNovita AI raised Deepseek V4 Flash 0731 token rates over 3x, added a new DeepSeek V4 Pro 0813 SKU, swapped an RTX GPU tier for the returning RTX 6000 Ada 48GB, and delisted most of its Image and Video model catalog.
DetailsPerplexity's developer docs renamed the Gateway API to Router API and added a private-preview banner requiring an email request for access, ending its self-serve availability. All 5 model prices are unchanged.
DetailsRegie.ai rebranded to RegieGO, replacing its $180-$499/user/mo annual sales-led seats with a self-serve Free plan ($0, 250 credits) and Pro at $49/mo or $41/mo billed annually (5,000 credits/mo); Enterprise stays custom-quoted.
DetailsSchematic replaced its two-meter pricing (monetized subscriptions + events) with one billing-volume meter and raised Growth from $200/mo to $400/mo.
DetailsAmbience Healthcare restructured its commercial model into performance-based contracts with fees tied to measurable clinical and financial outcomes, replacing its reported per-clinician license with forward-deployed teams accountable for results.
DetailsDeepSeek's confirmed time-of-use rate card took effect as scheduled on 2026-08-16, replacing flat per-token API rates with peak/off-peak pricing up to 12x higher, alongside a new experimental vision model on the same rate card.
DetailsAdobe added Firefly for Teams, a per-license Pro/Pro Plus/Premium business tier with admin tools and 24/7 support, priced the same as its Individual plans.
DetailsDeepL's Pro pricing page now lists a new 'Translation Flow' localization-workflow automation add-on on Business and Enterprise tiers only, with no price disclosed on the pricing page itself.
DetailsSarvam's docs pricing page carries a Dubbing product priced per minute of source audio (Starter ₹40, Pro ₹38, Enterprise ₹36), doubling under an editor_flow flag for the interactive Creator Studio workflow — a new priced SKU absent from the main marketing pricing page.
DetailsDeepSeek confirmed peak/off-peak API pricing effective 2026-08-16: peak rates run roughly 3-12x today's flat prices, off-peak rates roughly 1.5-2.5x.
DetailsFireworks added Muse Glimmer 30B and NVIDIA Nemotron 3.5 Lightning 30B A3B to its serverless rate card; Nemotron 3.5 Lightning's $0.05 input rate is now the cheapest published on the entire card.
DetailsFLORA's Special launch offer — extra included usage worth $12/seat (Starter), $50/seat (Pro), and $100/seat (Max) per month — had expired Aug 1 and disappeared from the pricing page, but reappeared on Aug 14 with a new deadline of Aug 31.
DetailsHebbia launched Max, a dealmaking AI agent for bankers and law firms, alongside Matrix. Like Matrix, it discloses zero pricing — access is gated behind a demo request.
DetailsHiggsfield's long-gated pricing page rendered live prices for the first time, revealing Starter/Plus/Ultra/Business replaced by Basic/Pro/Max and new per-seat Team/Scale plans, with no free tier shown.
DetailsThe 'Save 35%' Essential offer that returned 2026-07-21 is no longer showing on Intercom's pricing page or calculator; Essential is back to its $29/seat/mo annual ($39 monthly) list price with no discount banner. fin.ai/pricing's Intercom-helpdesk seat anchor still shows $19.
DetailsInVideo's live pricing page finally loaded real numbers after months of API failures: Plus $17/mo, Max $85/mo, Generative $170/mo and Elite $900/mo, plus Team seats from $40/mo — replacing prior third-party estimates of a free tier and roughly $25/$60 paid tiers.
DetailsLiveKit Inference gained one new LLM SKU, Google Gemini 3.7 Flash, priced at $0.0029/min, while all LiveKit Cloud plan prices (Build free, Ship $50, Scale $500, Enterprise custom) held unchanged.
DetailsModal now limits its token-billed Shared API (launched with Kimi K3) to Team and Enterprise plans; Starter keeps the same models via a per-second Auto Endpoint instead.
DetailsNetlify halved its Gemini 3.6 Flash token rate and added Gemini 3.7 Flash at the same price, cutting AI Gateway credit costs for teams using those Gemini models.
DetailsNovita AI delisted Ling 3.0 Tiny, the free LLM added 3 days earlier, leaving no $0 language model on its rate card as the catalog eased from 177 to 174 listings.
DetailsPerplexity's developer docs added an explicit end-of-support date to the existing Sonar-to-Agent-API migration banner, giving developers a firm deadline to move off the original Sonar Chat Completions API.
DetailsPerplexity added two NVIDIA Nemotron models to its week-old Gateway API — nemotron-3.5-lightning-30b-a3b, now the catalog's cheapest model at $0.0115/1M input tokens, and nemotron-3-ultra-550b-a55b — growing the Perplexity-hosted open-weight catalog from 3 to 5 models.
DetailsPerplexity's enterprise pricing page came back online after a month behind a Cloudflare wall, relaunching with annual-only per-seat pricing and a new rule: SCIM, audit logs, RBAC, and data retention now require a signed contract on both Enterprise tiers, replacing the old 50-seat threshold gate.
DetailsPerplexity's Agent API fetch_url tool price reverted to $0.0005 per invocation, undoing the halving to $0.00025 that took effect on 2026-07-29 — confirmed by the docs' own worked cost example, whose total rose from $0.00675 to $0.007.
DetailsSambaNova Cloud's Free plan no longer hands new developers $5 of API credit with no card required; it now asks users to add a payment method and purchase credits before making any requests.
DetailsSarvam AI's API docs pricing page now lists Sarvam-105B at ₹29.28/₹10.98/₹73.2 per 1M tokens, up from ₹4/₹2.5/₹16, and adds two new beta models (GLM-5.2, Gemma-4 31B) not on its marketing pricing page.
Detailsfal added Google's Veo 3.1 (4K, native audio, video extension) and repriced the whole Veo line: Standard drops to $0.20/s no-audio ($0.40/s with audio) from $0.50/$0.75, Fast to $0.10/s ($0.15/s) from $0.25/$0.40. The old Veo 3 and Veo 3 Fast endpoints are now deprecated.
DetailsFireworks published a forward-dated price increase on every on-demand GPU tier: H100/H200 from $7.00 to $8.00/hr (+14%), B200 from $10.00 to $13.00/hr (+30%), B300 from $12.00 to $15.00/hr (+25%), and GB300 from $18.00 to $20.00/hr (+11%), effective September 1, 2026. It is the first repricing of an already-published on-demand SKU since the product launched in January 2024.
DetailsMercor added three new proof points to its Enterprise page -- a 25-workflow diagnostic, a $20M by-2031 value-creation assessment, and a 132-run HR evaluation -- doubling its case-study count while buyer pricing stays undisclosed.
DetailsTogether AI added Kimi K3 as a third Provisioned Throughput model and raised per-PTU capacity 18-80% for MiniMax M3 and GLM-5.2 while holding the $0.05/PTU-minute sticker price flat.
DetailsWeights & Biases removed five models from its Serverless Inference rate card and added one new model, NVIDIA Nemotron 3.5 Lightning, at $0.10 in / $0.05 cached / $0.25 out per 1M tokens — no surviving model's price moved.
DetailsNvidia-backed Nscale agreed on 2026-07-30 to acquire Anyscale, the managed Ray platform, for $1.65B ahead of Nscale's own planned IPO. Deal is confirmed via an official Nscale press release, TechCrunch, and Latham & Watkins advisory notice. This is a continuity/rollup event (Nscale absorbing Anyscale's compute stack) rather than a wind-down; Anyscale's per-ACU pricing was unchanged at the time of the announcement.
DetailsCerebras Systems (wafer-scale AI inference cloud, Nvidia challenger) completed its IPO in the week of 2026-08-03, pricing above its initial range and reportedly 20x oversubscribed on debut, before shares fell over the following days. Confirmed via Stocktwits, The Motley Fool, TradingKey, and renaissancecapital.com. Cerebras's usage-based per-token API pricing (Free Trial, $10 Developer, Enterprise) plus the fixed-price Cerebras Code subscriptions were unchanged at capture time.
DetailsOVH Groupe (OVHcloud) completed its acquisition of speech-to-text/audio-intelligence API Gladia around 2026-07-31/08-03, issuing 1.8M OVH shares as part of consideration. Confirmed via an official OVH Groupe release (Idéal Investisseur) and TipRanks, both framing it as a 'sovereign voice AI' play. Gladia's prepaid-credit-wallet pricing ($0.61/hr async, $0.75/hr real-time, commit-based Growth from $0.20/hr) was unchanged at capture time — reads as a continuity/integration event, not a wind-down.
DetailsKlaviyo acquired Agency, an AI-focused startup founded by Elias Torres (Drift co-founder and a former Klaviyo advisor/mentor figure), around 2026-08-05, naming Torres Chief Product Officer. Widely and consistently covered (TechCrunch, Yahoo Finance, Business Journals, Dealroom, Wowtale) as an AI-capability acquisition, disclosed alongside Klaviyo reporting $370.6M revenue. No pricing-page change observed; Klaviyo's blueprint page is still a bootstrap stub pending its first full capture.
DetailsThe Wall Street Journal reported (via Yahoo Finance, 2026-07-23) that Stripe is in talks to acquire OpenRouter, the pass-through multi-model LLM routing marketplace, in a deal potentially worth ~$10B. Corroborated by Seeking Alpha and other outlets citing the same WSJ report. As of this note the deal is UNCONFIRMED/pending ("in talks", not announced or closed) — no pricing change has occurred; OpenRouter's zero-markup pass-through model plus 5.5% credit-purchase fee remains as last captured.
DetailsTricentis (enterprise software testing platform) agreed on 2026-07-30 to acquire Tabnine, the private/deployable-anywhere AI coding assistant, in a deal estimated at tens of millions of dollars. Confirmed via Business Wire, DevOps.com, ERP Today, and Israeli outlet calcalistech.com. Tabnine's per-seat pricing (Code Assistant $39/mo, Agentic $59/mo) was unchanged at the time of the announcement; strategic rationale is grounding Tricentis's testing agents in Tabnine's context/knowledge-graph technology, not a standalone product wind-down.
DetailsTempus AI agreed on 2026-07-21 to acquire cancer-monitoring/MRD platform Personalis (NASDAQ: PSNL) in a $1.5B cash-and-stock merger, expanding Tempus's genomics/data-licensing lines. Confirmed via Yahoo Finance, builtinchicago.org, and Personalis's own Q2 2026 earnings coverage (merger announced alongside results). Tempus has no public self-serve price sheet (genomics is payer-reimbursed, data/apps are quoted) so this is not expected to move a listed price, but it materially expands the scope of Tempus's data-licensing business line.
DetailsYellow.ai agreed on 2026-08-03 to go public via a SPAC merger with Bluerock Acquisition Corp (BLRK), valuing the combined company at ~$550M ($300M stock deal plus $5M equity / $50M note PIPE). Confirmed via an official PR Newswire release plus consistent coverage from Stock Titan, TradingView, Investing.com, and Unite.AI. No pricing change observed; this is a public-listing/capital-structure event, not a repricing.
DetailsAugment cut GPT-5.6 Luna token rates 80% ($1.00/$6.00 to $0.20/$1.20 per MTok) and Terra 20% ($2.50/$15.00 to $2.00/$12.00), adding Sonnet 5 and Opus 4.5.
DetailsChargebee replaced its Starter (free to $250K cumulative billing) and Performance ($7,188/yr) tiers with a single Flow plan billed at 0.80% of billing pay-as-you-go or 0.65% plus a $99/mo fee on commit, ending the free entry threshold.
DetailsCrewAI's free Basic plan drops from unlimited automations to a 2-per-month cap, and Enterprise's execution ceiling moves from Unlimited to Custom (quote-dependent).
DetailsCursor's docs now price GPT-5.6 Luna at $0.2 in / $1.2 out per 1M tokens (down 80% from $1/$6) and GPT-5.6 Terra at $2/$12 (down 20% from $2.5/$15) in the Other Models usage pool.
DetailsDeepInfra cut DeepSeek-V4-Flash-0731 pricing ~11%: $0.09/$0.18 to $0.08/$0.18 per 1M tokens (in/out), a second catalog-only model repriced outside the main /pricing page.
DetailsDeepSeek's Models & Pricing docs now carry a footnote warning of a coming, unspecified 'significant' overall price increase for its API; current per-token rates are unchanged as of this capture.
DetailsDiffbot relaunched its pricing page with a new Web Search API (1 credit per query) folded into the same all-APIs entitlement as Extract, Natural Language, and Knowledge Graph on every tier; headline plan prices are unchanged.
Detailsfal's Serverless & Compute table now lists the H100 80GB GPU at a $4.50/h List Price, up from $3.99/h three weeks earlier — a 12.8% rise that erases the price gap to the H200. The negotiated 'as low as' floor held at $1.89/h.
DetailsFireworks added GB300 288 GB to its on-demand GPU card at $18.00/hr and introduced a 1.5x region-restricted deployment premium for US/Europe-pinned dedicated GPUs, mirroring its existing 10% US-only Serverless token premium.
DetailsFLORA launched Fashion Studio, purpose-built fashion generation tools, and bundled it into all plans at no extra seat price — Free gets partial access, Starter/Pro/Max get full access, and usage draws from each plan's existing shared pool.
DetailsFLORA's Special launch offer, which added extra included usage worth $12/seat (Starter), $50/seat (Pro), and $100/seat (Max) per month, ran through Aug 1 and is no longer shown on the pricing page as of Aug 11.
DetailsGlean's Core Suite Model Hub Usage table added GPT-4o Mini TTS and Deepgram Nova-3 rates and split GPT Realtime pricing into per-modality lines, disclosing real-time audio token rates for the first time.
Detailsgroq.com/pricing/ now 308-redirects to Groq's homepage, dropping the dedicated marketing pricing page entirely; per-token, per-hour, and per-character rates are unchanged but now live only inside the GroqCloud developer docs model catalog.
DetailsHyperbolic's dedicated Serverless Inference pricing page now 404s; its docs show only two priced models ($0.40/M each) and mark the entire image-model catalog plus several text/vision models for sunset.
DetailsInstantly raised the Outreach Hypergrowth plan Monthly-billing email allowance from 100,000 to 125,000 emails per month at the same $97/mo price -- a 25% capacity increase that the Yearly toggle and feature-comparison table do not reflect.
DetailsInVideo's invideo.io/ai/pricing now returns a server error, leaving invideo.io/pricing as the sole pricing surface, which now splits plans behind Individual and Team & Enterprise tabs — though exact prices still fail to load on both tabs via a persistent client-side API error.
DetailsLindy moved to per-seat billing with a shared team credit pool, cut Plus from $49.99 to $29.99/user/month, and dropped its $8,000/month human-assistant anchor.
DetailsLiveKit's per-minute LLM pricing dropped for two existing GPT-5.6 Inference SKUs: Luna fell 80% to $0.0008/min and Terra fell 20% to $0.0081/min, while Sol held at $0.0203/min and all LiveKit Cloud plan prices stayed unchanged.
DetailsMake dropped the three named paid tiers it introduced in June 2026 (Core, Pro, Teams) and returned to one 'Make Plan' at $9/mo annual — but the entry credit step halved from 10,000 to 5,000 credits/mo, an effective price increase per credit at the same headline price.
DetailsMem0's Pro and Enterprise plans gained a new 'Dream (Memory Consolidation)' capability, and the pricing page footer now shows HIPAA Ready, SOC 2 Type I, and GDPR Ready badges. Tier prices and quotas are unchanged.
DetailsMercor rebuilt its Enterprise page, dropping the $10B/$2B+/100k+/400+ scale-stats bar and self-referential AI-support case study for three named client engagements, and renamed its Agent Benchmarking service to Agent Optimization under a new Discover-Deploy-Improve-Monetize framing.
DetailsMintlify reintroduced a paid Pro tier ($450-540/mo) and moved its AI assistant and automations off the free Starter plan, which now ships only the docs platform.
DetailsModal raised its Team plan's included container concurrency from 1,000 to 5,000 (a 5x increase) with no change to the $250/mo + compute plan fee, while also itemizing new Team-tier features: custom domains, static IP proxy, deployment rollbacks, and RBAC.
DetailsNomic added a self-serve Free tier and a $20/month Individual tier to its AEC Nomic Platform, opening a PLG entry point below the $40/user/mo Business plan (25-seat minimum), and dropped the previously stated $1,000/mo platform-commitment line from the pricing page.
DetailsMacaron V1 Venti, Macaron V1 Tall, and Ling 3.0 Flash moved off Novita AI's introductory $0 rate to paid per-token pricing, while a new model, Ling 3.0 Tiny, took over the free-launch slot.
DetailsOne week after restoring its two-tab Regular/Enterprise layout, Oxylabs flattened the Web Scraper API pricing page back to a single four-column grid; per-1K rates on the surviving tiers are unchanged.
DetailsPerplexity added a fifth developer API, the Gateway API, offering unified access to open-weight models it hosts itself (DeepSeek, Kimi, GLM) at Perplexity-set per-token prices from $0.13/1M input tokens, distinct from the Agent API's at-cost third-party resale.
DetailsPinecone's Nexus agent knowledge engine moved from public preview to general availability on the pricing page. Access remains request/trial-gated with no public price disclosed; core Database, Inference, and Assistant rates are unchanged.
DetailsSambaNova Cloud reversed its July 2026 cut to gemma-4-31B-it, raising the rate back to $0.38 input / $1.15 output per 1M tokens from $0.22/$0.59, so gpt-oss-120b is now the sole cheapest model on the rate card.
DetailsUpstash removed the 'Upstash Box is in Developer Preview! APIs and pricing may change' disclaimer from its Box pricing page and folded Keep Alive into a per-box toggle inside Pay as You Go, with no price changes across any of its six products.
Detailsv0 by Vercel cut per-token rates for its Mini model by about 80% ($1/$5 to $0.20/$1.20 per 1M in/out) and Pro by about 33% ($3/$15 to $2/$10), while Max and Max Fast and all seat prices stayed flat.
Detailsv0 Mini tokens drop ~80% and v0 Pro tokens ~33%; AI Gateway adds a metered Trace Drains SKU and Enterprise invoice billing; Custom Environments gets published overage pricing; Sandbox concurrency quota rises from 2,000 to 10,000.
DetailsW&B cut DeepSeek V4-Pro Serverless Inference rates roughly 34% on input and 27% on output, and added a new model, DeepSeek V4-Flash-0731, at $0.13 in / $0.28 out per 1M tokens.
DetailsWellSaid's free Trial dropped its 7-day time limit and now includes 3 downloaded minutes/month (no commercial rights); Starter and Pro gained a new 'Extra minutes available' self-serve top-up, addressing the plan's prior no-overage gap.
DetailsThe help-center free student plan that appeared around 2026-07-30 (unlimited words, no credit card) has been withdrawn — the Flow Discounts article is back to describing only the standard discounted-Pro student offer.
DetailsWispr shipped a second product, Wispr Notetaker (meeting transcription, speaker ID, cross-meeting Q&A, MCP export), as a new nav item and help-center collection — included in Free and Pro at no extra seat price, with Enterprise marked 'coming soon'.
DetailsYou.com added a new Answer API that returns a synthesized, citation-grounded answer in one call, priced at $5.00 per 1,000 calls — the same rate as its raw-results Search API — with 93.48% claimed accuracy on SimpleQA.
DetailsFramer introduced monthly AI-credit pools (500/1,000/3,000 credits) for its Agents features and removed the intermediate Scale plan, leaving Free / Basic / Pro / Enterprise site plans.
DetailsAleph Alpha's homepage no longer mentions 'PhariaAI' anywhere, pivoting to generic 'Sovereign AI' and specialized-language-model (SLLM) messaging; pharia.com and pharia.ai now redirect to the plain homepage instead of a standalone product site.
DetailsCursor repriced two third-party models in its Other Models token pool: GPT-5.6 Luna fell about 80% and GPT-5.6 Terra about 20% across input, cache, and output rates, while GPT-5.6 Sol held steady.
DetailsDeepSeek's pricing docs now list a Responses API feature (Flash-only today) and a footnoted plan to charge 2x during Beijing-time peak hours, though no effective date or per-token price has changed yet.
DetailsMagnific's pricing page now gates Business and Enterprise behind a separate Teams tab, and the Business plan gains admin controls, no-training-on-your-data, and Magnific Studios access that were previously Enterprise-exclusive.
DetailsHebbia's site now lists Max, a deal-staffing AI assistant for banking, PE and legal, as a distinct nav item next to Matrix. No pricing is published for Max, and Hebbia's existing sales-only per-seat model for Matrix is unchanged.
DetailsMake replaced its three-tier Core/Pro/Teams ladder with a single all-inclusive paid tier, reverting the credit-slider entry step from 10,000 back to 5,000 credits/mo at the same $9/mo annual price.
DetailsMem0's comparison table now lists a new 'Dream (Memory Consolidation)' capability checked on Pro ($249/mo) and Enterprise only, and the site footer added SOC 2 Type I and HIPAA Ready compliance badges; no prices or quotas changed.
DetailsA new site-wide banner on Nomic's pricing and product pages reads 'Self Service Starts At $20/Month,' the first sign of an entry tier below the existing 25-seat-minimum Business plan, though plan details are not yet published.
DetailsSarvam AI docs pricing page now shows Sarvam-105B priced ~7x higher (Rs 29.28/10.98/73.2 per 1M tokens) plus two new third-party models; marketing page unchanged.
DetailsDiffbot launched a Web Search API (1 credit per search) alongside Extract, Crawl, Natural Language, and Knowledge Graph, folding it into the existing $0/$299/$899/custom credit-pool plans without changing any headline price.
DetailsModal raised its Team plan's container concurrency limit from 1,000 to 5,000 (5x) at the unchanged $250/month plus compute price, boosting included capacity without touching the rate card.
DetailsPerplexity quietly launched a Gateway API hosting open-weight models (Moonshot AI's Kimi K3, Z.AI's GLM 5.2) at its own per-token rates, a fifth metered developer surface distinct from the Agent API's at-cost third-party resale.
DetailsBaseten added DeepSeek V4 Flash and Inkling-Small to its Model APIs catalog (now 12 models), renaming DeepSeek V4 to DeepSeek V4 Pro at unchanged rates.
DetailsDeepInfra cut GLM-5.2 pricing ~20%: $0.93/$3.00 to $0.75/$2.40 per 1M tokens (in/out), while every other featured catalog model held steady.
DetailsFrase's pricing page now shows three explicit FraseCMS hosting tiers -- Free (100K page views), Basic $19/mo (250K page views, custom domain), and Pro $99/mo (1M page views, 500GB bandwidth) -- stacked on top of any paid plan, where FraseCMS was previously an unpriced bundled perk.
DetailsGlean's Model Hub Usage rate card cut GPT 5.6 Terra 20% and GPT 5.6 Luna 80% per million tokens, and reclassified Luna from Premium to Standard in the Enterprise Flex FlexCredit tier table.
DetailsHubSpot extended Revenue Hub's platform-fee waiver from 60 days to 1 year, a 6x jump in the same promo shown on its payments page just five days earlier.
DetailsHyperbolic raised H100 SXM to $3.19/GPU/hr (+10%) and H200 to $3.99 (+14%); its 'starting at' price now equals the H100 rate, replacing the old $0.20/GPU/hr floor.
DetailsHyperline's flat $599/month Launch tier lasted about a week: Launch is now $299/month plus 0.7% of revenue processed (still up to $2M revenue, 100 active subscriptions). Growth and Scale remain custom-quoted.
DetailsIdeogram's API added a P-Image model (US$0.003-US$0.03/img) plus Object Removal and Describe V4 endpoints; consumer plan prices unchanged.
DetailsMetronome's Starter plan bullets no longer state an included billing-volume or event allotment; it now reads only the flat metered rates (0.8% of billing volume, $0.04/1,000 events), three weeks after first publishing a $100,000-plus-10M-event allotment.
DetailsMintlify added a $450-$540/mo Pro tier between Starter and Enterprise, moving AI agent features and its per-credit meter off the free Starter plan entirely.
DetailsMoonshot pulled Kimi Enterprise's only public price — live through July 30 — moving the tier to a sales-only 'Contact us' motion with no rate disclosed.
DetailsNetlify's pricing page no longer shows a starting price for Enterprise -- the card now reads only Custom instead of the previously disclosed Custom, starts at $500/month, aligning it with the fully sales-quoted convention most contact-sales tiers use.
DetailsNovita AI removed the RTX 6000 Ada 48GB instance from its self-serve GPU lineup and added a new base RTX 5090 32GB on-demand tier at $0.73/hr, while its model catalog grew from 172 to 177 listings.
DetailsOxylabs merged its Dedicated Datacenter Proxies product page into the general Datacenter Proxies page via a 301 redirect; the $2.25/IP entry tier now appears as a 'Dedicated (pay per IP)' card there, with the free trial demoted from a plan card to a banner.
DetailsOne month after flattening its Web Scraper API pricing page into a single grid, Oxylabs restored the two-tab Regular/Enterprise layout with all eight named tiers back; every per-1K rate is unchanged.
DetailsPoolside cut its Laguna S 2.1 API rate 10% to $0.09/$0.18 per 1M input/output tokens on OpenRouter, the first price move since the model's $0.10/$0.20 July 2026 launch; Laguna XS 2.1 holds at its existing 40%-off $0.06/$0.12 rate.
DetailsSpeechmatics replaced the recurring 3,000-minute (50-hour) plus 1M-character monthly free allowance with a one-time $100 usage credit for new accounts, effective around 31 July 2026.
DetailsA help-center article now describes a standalone free, unlimited-words student plan claimed via a desktop-app sidebar card, alongside the existing 3-months-free-then-$6/mo Pro discount still shown on the public students page.
DetailsFlow Studio adds two new workspaces, a browser-based 3D Editor and a node-based Canvas for image/video generation, with no published credit cost or tier gating yet and no change to the Free/Lite/Standard/Pro/Enterprise rate card set in August 2025.
DetailsxAI added a pricier Speech-to-Speech voice model at $0.08/min ($4.80/hr), 60% above the existing model's $0.05/min ($3.00/hr) rate, and renamed the Voice API line item.
DetailsZenRows replaced its dollar CPM/per-GB pricing with a shared credits currency across new Fetch/Extract/Batch/Browser Sessions primitives, and made Free a permanent $0 plan instead of a 14-day trial.
Details01.AI is winding down its yi-lightning / yi-vision-v2 API: new signups and top-ups stopped immediately, API access ends 2026-09-03, and refund requests are accepted through 2026-12-03.
DetailsAI Gateway now offers Trace Drains, forwarding an OpenTelemetry trace of every request to a customer's own observability tool for $0.05 per 1,000 traces plus $0.50 per GB of trace egress on Pro and Enterprise.
DetailsBraintrust renamed its 'Topics credit' to a shared 'Model credits' pool and quietly dropped GLM-5.2's July 31, 2026 sunset date, making it a standing feature.
DetailsCrewAI's Enterprise plan drops its bundled 50 dev-hours/month for a 45-day onboarding program; forward-deployed engineering and training are now a la carte.
DetailsFirecrawl now grants every plan, from the free tier to custom Enterprise, an identical +1,000 search credits per month on top of its regular credit pool, timed with a pricing-page banner touting its most accurate /search yet.
DetailsGladia added a plan comparison table showing a 99.9% uptime SLA and priority queue now included on Growth and Enterprise; per-hour rates are unchanged.
DetailsGoogle launched Gemini Robotics ER 2 Preview and ER 2 Streaming Preview, embodied-reasoning models priced at $2.00/1M input and $10.00/1M output, double the existing ER 1.6 Preview rate, alongside four new Gemma 4 fine-tuning SKUs on Vertex AI.
DetailsGumloop removed its long-standing $0/mo, 5,000-credit Free plan and now leads with a 14-day free trial of Pro instead, offered only on monthly billing. The Pro slider curve itself ($37/20k credits up to $1,840/1M credits monthly, 20% off annually) is unchanged.
DetailsHubSpot shipped Revenue Hub, a CPQ, billing and payments product priced Free / from $95 / from $140 per month, with card payments carrying a 2.9%+0.5% transaction fee.
DetailsHubSpot renamed its Breeze AI agents to Agent Hub (Beta) and added HubSpot AEO and Agent Builder, while per-outcome prices stayed unchanged at $0.50, $1.00, and $0.10.
DetailsMiniMax launched MiniMax-H3, an open-weight omni-modal video model that replaced Hailuo as its video product, moving video billing from $0.19-$0.56 per clip to $0.08-$0.13 per second.
DetailsPoolside now sells open-weight Laguna models at $0.06-$0.20 per 1M tokens via OpenRouter, its first public price after years of sales-only enterprise deals.
DetailsApify doubled per-Actor RAM allocations across Free, Starter, Scale, and Business at unchanged prices (e.g. Free 8GB to 16GB, Business 256GB to 512GB), and dropped the standalone $100 priority-support add-on from its pricing page.
DetailsArize removed the explicit $10-per-million-spans and $3-per-GB overage rates from its pricing page; headline prices ($0 Free, $50 Pro) and included quotas are unchanged, but overage pricing is no longer publicly disclosed.
DetailsBaseten grew its Model APIs catalog to ten models with new flagship Kimi K3 and GLM-5.2 Fast, and cut GLM-5.2's cache-input rate 46% to $0.14 per 1M tokens.
DetailsAnthropic introduced Claude Opus 5 as the featured flagship model on claude.com/pricing, at the same $5/$25 per-MTok input/output rate as Opus 4.8, which now moves to the page's Legacy models list.
DetailsDeepInfra raised Llama-3.1-8B-Instruct-Turbo output price 33% to $0.04 per 1M tokens (from $0.03) and cut gemma-4-31B-it-turbo to $0.09/$0.34 per 1M.
DetailsExa raised free-tier credits (signup $10 to $20, monthly $7 to $10), dropped the card requirement, and replaced its 20,000-req/month framing with pure credits.
DetailsFireworks added Kimi K3 (plus Fast and a 10%-premium US-only variant) to its serverless rate card, shipped a new Serverless Training API for per-token LoRA training, and switched Fire Pass's free model to Kimi K3 Fast with a 1M-token context window.
DetailsHedra merged its Hedra Avatar and Hedra Omnia native models into a single Character-3 model at 6 credits/second and replaced roughly a dozen third-party Kling/Seedance/VEED video models with a new MiniMax Hailuo 2.3 and Hailuo-02 family.
DetailsHyperline replaced its $199/$299 + 0.6%/0.7%-of-revenue pricing with flat tiers: Launch at $599/month (up to $2M revenue, 100 contracts), plus custom-quoted Growth and Scale bands. The 10-invoices-free trial is also gone.
DetailsKrisp's Voice Translation API moved from a 'Soon' badge to self-serve: 60 minutes free, Python/JS SDKs, no sales call needed — inside an otherwise contact-gated Voice AI SDK.
DetailsLangSmith's pricing page now checks Email Support for Enterprise only; the Plus plan drops direct email support in favor of a new self-serve Support Portal, while Enterprise also gains a new Slack support channel.
DetailsLinear renamed 'Linear Agent automations' to 'Loops' and now meters it with AI credits, a second metered AI feature alongside Coding Sessions on its per-seat plans.
DetailsLiveKit Inference gained a new TTS vendor, Fish Audio (S2 Pro, S2.1 Pro, and a free S2.1 Pro Free SKU), plus Google Gemini 3.5 Flash Lite and Gemini 3.6 Flash LLM entries. Cloud plan prices are unchanged.
DetailsMercor's Data page now advertises 'Earn up to $100K per referral' for enterprises referred into its data-partnership program, replacing a prior undisclosed-amount referral promise -- the first concrete dollar figure Mercor has shown on that surface.
DetailsModal added a new OpenAI-compatible Shared API billed by token rather than GPU-second, launching first with Moonshot's Kimi K3 model; per-token rates are not yet published on Modal's pricing page.
DetailsNovita AI's Flux.1 Kontext Pro image price jumped 10x, from $0.036 to $0.36/image, and its model catalog shrank 196 to 172 as legacy video SKUs were delisted.
DetailsPerplexity cut its Agent API web_search tool price in half to $0.0025 per invocation and fetch_url to $0.00025, and applied the same cut to SDK search queries run inside its sandbox tool.
DetailsPlayground cut Free tier limits: downloads fell 10→2/day and its rolling-window generation cap fell 10→5 images/3h. Pro and Pro Plus pricing is unchanged.
DetailsResemble AI added Team ($280/mo, 5 seats) and Business ($800/mo, 20 seats) tiers with discounted usage rates, reversing June's move to pure pay-as-you-go Flex pricing.
DetailsSanctuary AI dropped Phoenix/Carbon branding for a 'Physical AI' platform spanning robotic arms, hands and humanoids — still sales-led with no published price.
DetailsSocket quietly expanded its Free plan's tracked-dependency cap from 1,000 to 10,000 -- a 10x increase -- while Team ($25/dev/mo) and Business ($50/dev/mo) list prices held steady.
DetailsStripe restructured Radar's fraud pricing into a four-tier Standard/Plus/Pro/Lite ladder, raising rates from ₹4-₹6 to ₹5-₹8.50 per transaction plus a new Pro per-customer fee.
DetailsSurfer SEO now sells AI Search Analytics as its own top-level product tab, priced by a 50/100/200-prompts-tracked-per-day slider ($82-$308/mo annual, unlimited custom-quoted), separate from its $49-$299/mo Full AI SEO content-optimization ladder, which is unchanged.
DetailsTavus removed the introductory discount it had just added to its Builder developer plan on July 21 — the tier now shows a plain $59/mo, and no tier on the pricing page carries a promo badge.
DetailsTogether AI raised GPU cluster reserved H100 rates and unified speech-to-text pricing at $0.0015/audio-minute, cutting Nemotron 3.5 ASR pricing 67%.
DetailsVellum replaced its single $50/mo Pro configurator with three preset packages (Mighty $30, Super $100, Ultra $200/mo) plus a Custom build option from $50/mo.
DetailsW&B cut Serverless Inference rates up to 45% on GLM 5.2, Kimi K2.7/K2.6, and Gemma 4 31B; raised GPT OSS 120B output 21%; and added MiniMax M3 to the catalog.
DetailsYou.com's Web Search API now gives 100 free calls per day before its $5-per-1,000-calls rate applies, replacing the prior flat meter; the docs haven't caught up.
DetailsAnthropic launched Claude Opus 5 (model ID claude-opus-5) as its recommended default for agentic coding and enterprise work, priced identically to Opus 4.8 at $5/$25 per 1M tokens — a model refresh, not a price move.
DetailsAnthropic lowered the Claude.ai Team plan's minimum team size from 5 seats to 2, widening self-serve access to small teams while leaving all seat prices ($20/$25 Standard, $100/$125 Premium) unchanged.
DetailsGoogle added two new Gemini API/Vertex AI models (3.6 Flash at $1.50/$7.50 per 1M tokens, 3.5 Flash-Lite at $0.30/$2.50) plus a new Gemini Spark agent for AI Ultra, while a US-region capture confirmed Google AI Plus at $4.99/month for the first time.
DetailsHugging Face added AWS NVIDIA H100 ($4.50/hr) plus its first Blackwell GPUs, B200 ($9.25/hr) and RTX PRO 6000 ($2.75/hr), to Inference Endpoints pricing.
DetailsPoe's $19.99/mo Premium plan now grants 660,000 compute points per month, down from 1,000,000 — the sticker price is unchanged, but included usage shrank by roughly a third.
DetailsAssemblyAI added a seventh priced product: a Sync Speech-to-Text API at $0.45/hr that returns finished transcripts in a single synchronous call with ~134 ms p50 latency and Universal-3.5 Pro accuracy.
DetailsCaptions dropped its $9.99/mo Pro plan; the paid lineup is now Max ($24.99), Scale ($69.99-$279.99, 1x/2x/4x usage), and Enterprise, with a free tier for editing and captions.
DetailsAnthropic expanded Claude Pro, Max and Team to include Claude Cowork, Design and Science alongside Claude Code at unchanged prices, and launched Claude Sonnet 5 on the API at an introductory $2/$10 per MTok.
DetailsComposio will replace its single tool-call meter with a six-dimension usage model on Aug 15, 2026: Free/Pro $29/Business $599/Enterprise. Pro tool calls drop 200K to 50K, overage rises to $4/1K.
DetailsDust's credit docs now show three action-credit tiers instead of two: a new Free (0-credit) tier explicitly covers memory, agent routing, common utilities, most orchestration helpers, skills management, and most Pods/Files actions.
Detailsfal's Serverless & Compute table now shows two hourly rates per GPU — a List Price and an 'As low as' rate. H100 lists at $3.99/h against the $1.89/h headline. B300 and RTX PRO 6000 join; A100 is gone.
DetailsKaiber's Help Center now says credits don't expire and roll over month to month, reversing the earlier cycle-reset forfeiture, and rebrands its batch/auto-clip product Cuts to Beat Sync. Plan prices held steady.
DetailsLlamaIndex raised included seats to 100 users on its Free, Starter, and Pro plans — up from 1 / 5 / 10 on the plan cards — while keeping prices, credit allotments, and the $1.25-per-1K credit rate unchanged.
DetailsLuma AI now prints an explicit monthly credit allowance on each Dream Machine Individual plan — Plus 10,000, Pro 40,000, Ultra 150,000 credits — ending its earlier usage-multiplier-only framing. Prices are unchanged.
DetailsMaxio's pricing page slimmed from three plans to two: Grow ($599/mo, up to $100k billings) and quote-only Scale. The free 'Build / Try for free' developer-sandbox column was removed from the compare table.
DetailsMiniMax's API card now leads with M3 and M2.7 at $0.30/M, launches Music 3.0 at $0.15/track, and adds per-call MCP and web_search billing — API-vlm cut to $0.01/call effective July 22, 2026.
DetailsOpenAI's API pricing page now leads with the GPT-5.6 flagship line: Sol ($5/$30 per 1M), Terra ($2.50/$15), and Luna ($1/$6), replacing the GPT-5.5/5.4 headline lineup and repricing the everyday tier.
DetailsPinecone Nexus, a knowledge engine for AI agents that offloads retrieval, assembly, and reasoning from the LLM to a dedicated layer, entered public preview. Access is request-only with no public price yet.
DetailsOn July 8, 2026 SambaNova announced the first close of a $1B financing round at an $11B valuation, led by General Atlantic, with JPMorganChase named as a new RDU inference customer.
DetailsSambaNova Cloud trimmed its rate card to six models, cut gemma-4-31B-it from $0.38/$1.15 to $0.22/$0.59, and added a Cached Input Tokens column led by MiniMax-M2.7 at $0.06/1M cached.
DetailsSarvam repackaged its prepaid credit plans: Pro bonus rose to Rs 2,000 (12,000 credits) and Business to Rs 12,500 (62,500 credits), and the Most Popular badge moved from Pro to the Rs 50,000 Business tier. Per-unit API rates were unchanged.
DetailsUsage AI's mid-2026 redesign removed the inline fee calculator and ICR-vs-ESR comparison from its pricing page; the fee is now described only as 'a percentage of realized savings,' with 20% (EC2) / 35% (RDS-class) surviving only on AWS Marketplace.
Details01.AI shipped TrueNorth in July 2026 — an enterprise AI decision platform packaged as Boss AI, Investor AI and TopSales AI, sold entirely by consultation with no published price.
DetailsAbridge's site now advertises an unlimited enterprise agreement (CHRISTUS Health) beside the per-clinician default, consolidates to three product lines, and claims 300+ health systems.
DetailsAI21 now sells Maestro as an optimization framework that cuts what production AI agents cost, adding per-team, per-repo and per-agent spend attribution. Jamba token rates unchanged.
DetailsBraintrust's pricing page now publishes a retention-storage rate — Pro keeps 30-day included retention, then bills $0.50 per GB per month for data held beyond it. Tier fees unchanged.
DetailsCartesia removed the Monthly/Yearly billing toggle from its pricing page, so self-serve tiers now advertise monthly-only rates: Pro $5, Startup $49, Scale $299.
DetailsChargebee's pricing page now runs four product tabs: Billing, CPQ, RevRec and Growth. Retention and Receivables are gone; CPQ Lite is free for 50 quotes.
DetailsFactory's custom-quoted mid tier is now called Business and gains audit logging plus network policy, both previously Enterprise-only. Seat prices unchanged.
DetailsFathom's pricing page gained a fourth paid tier. Seat prices are unchanged, but single sign-on left Team and custom data retention left Business — both are now Enterprise-only.
DetailsFireworks AI's site banner now announces a Series D round alongside a $1B annual recurring revenue milestone, four years after the company was founded in 2022.
DetailsFireworks launched Fire Pass, a promo-code pass with zero per-token charges for personal agentic coding, and now prices serverless on Standard, Priority and Fast paths.
DetailsFreepik's pricing page, now Magnific-branded, drops the entry Essential tier, raises Premium+ to $45/mo and Pro to $280/mo, and adds a $69/seat Business plan with a shared credit pool.
DetailsGladia moved self-serve billing to a prepaid credit wallet with auto top-up and swapped its recurring 10 free hours per month for a single non-resetting 50 EUR signup credit.
DetailsGlean's Enterprise Flex docs now meter 13 capabilities in FlexCredits, adding Adaptive Reasoning, Voice Sessions, per-minute Meeting Notes, Code Writer, Image Generation and Client API calls.
DetailsHeyGen now publishes Business credit top-ups at $0.05 per credit and avatar add-ons at $29/mo per slot, and raises Video Agent to 30 credits per minute.
DetailsIdeogram added its 4.0 model to the consumer credit matrix priced per single image (2/4/6 credits) instead of per four images, so a bundled allowance buys about 4x fewer 4.0 generations.
DetailsInflection AI's enterprise surface is gone: inflection.ai/enterprise returns a 404, Enterprise is absent from nav and footer, and developers.inflection.ai no longer resolves.
DetailsInflection AI added a Labs surface at inflection.ai/labs shipping Pi Journeys plus two 'Coming soon' cards, and put it in the top nav alongside Try Pi. None of it is priced.
DetailsJasper layered credit metering onto its per-seat plans: GEO Hub, Grid rows and heavier Agents now consume credits on the Business plan, per a new rate card.
DetailsLago replaced its two quote-only tiers with one Lago Premium package, moved self-hosting into the base offer, and unbundled 24x7 support and an assigned solution engineer as add-ons.
Detailsn8n replaced the per-plan AI Workflow Builder credit line with AI Assistant credits, lifting Starter from 50 to 2,300 credits a month and Pro from 150 to up to 13,700. Plan prices unchanged.
DetailsObserve.AI now markets Agent Platform, Agent Harness, CoBuilder, Interaction Intelligence and Interaction Fabric as separate platform modules under a new top-level Platform menu.
DetailsObserve.AI's AWS Marketplace VoiceAI Agents listing now prices two usage dimensions — $4.80 per Minutes-per-Month unit and $12.00 per Interactions unit for 12 months.
DetailsPhotoroom repackaged its app plans: the Free card is gone, and Pro/Max/Ultra now publish hard numbers - up to 8,000/25,000/75,000 AI credits and 1,000/3,000/10,000 exports per month.
DetailsRunPod published a dedicated Enterprise page pairing reserved capacity at committed-use rates with usage-based burst on a single post-paid invoice, plus Runpod Overdrive sold on a pay-nothing-if-we-miss guarantee.
DetailsStripe moved advanced usage-based billing out of its 0.7% Billing rate into Metronome, a Stripe product with its own Startup and Custom plans; basic metering stays at 100M events/mo.
DetailsSynthesia now quotes Indian buyers in rupees — Starter ₹1,999/month (₹1,499/mo yearly) — instead of the discounted $19/$14 USD tier it served that market in May.
DetailsVEED lifted its yearly-billed USD seat prices to $12, $22 and $39 per user per month — Creator up 20%, Studio up 11% — while leaving AI-credit allowances unchanged.
DetailsWispr Flow's pricing page now renders local currency by visitor country. From an India IP, Flow Pro shows Rs 320/user/mo annual and Rs 400 monthly against a $12/$15 US list.
DetailsYou.com replaced the open-ended greater-than-$2,000 label on its Frontier research-effort tier with a firm $1,200.00 per 1,000 calls, cutting the top of its Research API range by roughly 40%.
DetailsZ.ai's GLM Coding Plan now lists at $18 (Lite), $72 (Pro) and $160 (Max) per month with a monthly/quarterly/yearly discount ladder, up from roughly $10/$30/$80.
DetailsBaseten cut its Model APIs rate card from 11 models to 8, retiring GLM 5.1, GLM 5, Kimi K2.5 and Nemotron 3 Super, and added Inkling at $1.00/$4.05 per 1M tokens.
DetailsBland's per-minute voice rate now covers only LLM, STT and TTS. Telephony moved out of the bundle and is billed separately at carrier or pass-through cost.
DetailsCerebras turned its open Free tier into a $5-credit Free Trial and added Gemma 4 31B at $0.99/$1.49 per 1M tokens, while dating GLM 4.7 for deprecation.
Detailscodeium.com/pricing and windsurf.com/pricing both resolve to devin.ai/pricing. Free $0, Pro $20/mo, Max $200/mo, Teams $80/mo + $40 per full dev seat, Enterprise custom.
DetailsDevin's prices held at $0/$20/$200 and $80 + $40 per full dev seat, but the free tier is now a downloadable Devin Desktop editor with cloud agents moved behind Pro.
DetailsCoreWeave added two spot-only Standard Memory CPU shapes, withdrew the $15,000/mo 100G Virtual Direct Connect tier, and dropped its reserved-storage discount.
DetailsDeepInfra added a third per-request service tier: Flex bills 0.8x base price for non-production and asynchronous work, joining Standard (1x) and Priority (1.5x).
DetailsFreshworks rebuilt the Freshsales pricing page: the $0 three-user Free plan is gone, and the Freddy AI Agent add-on is now listed at $49 per 100 bot sessions.
DetailsGroq removed Llama 4 Scout and Qwen3 32B from its public rate card and docs catalog, dropped the $0.08/hour Browser Automation tool, and swapped enterprise-only Minimax M2.5 for M2.7. Retained model prices held stable.
DetailsHyperbolic now packages compute as On-Demand, Reserved, Private Cloud and Serverless Inference, funded by never-expiring prepaid credits with a $5 minimum and Auto Top-Up.
DetailsHyperbolic nearly doubled its published H100 SXM starting rate to $2.89/GPU/hr, raised H200 to $3.49 and B200 to $5.99, and pulled its consumer-GPU per-hour rates.
DetailsIntercom's new-customer Essential discount is back on both cadences: $29 struck to $19/seat/mo annual (Save 35%) and $39 struck to $29/seat/mo monthly.
DetailsJuicebox drops its Starter seat from $139 to $119/mo ($99 billed yearly), doubles Starter credits to 500, lifts its 3-project cap, and raises Growth to 1,500 credits per seat.
DetailsLangSmith dropped per-run and uptime deployment billing for two normalized units: LangChain Compute Units at $1.50 and new Storage Units at $1.00. Seats hold at $39.
DetailsLangSmith replaced its per-1k trace dollar rates with normalized LCU/LSU units. Base traces now cost 0.005 LSU each — double the old $2.50/1k — and extended retention drops to 180 days.
DetailsLiveKit Inference added six new per-minute LLM SKUs: OpenAI GPT-5.6 Luna, Sol and Terra, xAI Grok 4.3 and 4.5, and Moonshot Kimi K2.6. Cloud plan prices unchanged.
DetailsLorikeet capped its voice-resolution meter: the 1.50/1.20 credit rate now applies only to resolutions up to 3 minutes long. List prices are unchanged.
DetailsLovable now bills building, Cloud hosting and AI-gateway calls from a single credit balance. Cloud and AI dollar grants became credits: 20 Cloud and 4 AI credits per month.
DetailsSalesforce's investment in m3ter converted into ownership: the site now reads 'm3ter is now part of Salesforce'. The four-component gated quote is unchanged.
DetailsMem0 dropped its $79/month Growth tier roughly two months after launching it, leaving free Hobby, $19 Starter, $249 Pro and custom Enterprise on a rebuilt pricing page.
DetailsMercor's enterprise product lines were renamed site-wide from Business / Enterprise evals / Data partnerships to Enterprise agents / Human data / Data monetization, and the data-partnership surface grew from 34+ to 50+ integrations.
DetailsZilliz Cloud launched On-demand Compute in public preview — per-minute Query and Indexing CU billing for lake-scale search on external data — and now publishes entry prices for its Standard and Enterprise plans.
DetailsMoonshot's pricing docs now flag the legacy moonshot-v1 family — the models that priced the same weights by context window — for full platform sunset on August 31, retiring context length as a billing axis.
DetailsNovita AI listed Kimi K3 at $3/M input and $15/M output, moved Tencent Hy3 off free pricing to $0.14/$0.58 per M, and cut its model catalog from 221 to 196 SKUs.
DetailsPerplexity's Agent API added a $0.03-per-session code sandbox tool, its first meter that counts neither tokens nor requests, and widened third-party model resale.
DetailsPuzzle dropped the '$20k transaction volume' free-tier trigger for a footnote reading 'Starter is free for 2 months', and relabelled Starter/Core AI credits as lifetime rather than recurring.
DetailsRoboflow Core holds at $79/mo billed annually but now includes 360 credits/year instead of 180, and the plan card gained a dropdown selling larger included-credit packs for $120-$590 extra per month.
DetailsSequence launched an AR agent, agent guardrails and Sequence MCP, and rebranded receivables as AR Collections. Tier prices are unchanged: Growth stays $799 per month.
DetailsTavus pulled published pricing for PALs, its consumer AI-companion product. tavus.io/pals/pricing now 301-redirects to the developer pricing page and no Tavus surface lists the consumer plans.
DetailsTavus's Most Popular Builder plan now shows $59/mo struck through against a $29 first month, while the $22 Starter plan loses its discount and is relabelled 'No pay as you go coverage'.
DetailsTogether AI cut Kokoro-82M TTS from $10.00 to $4.00 per 1M characters and added Inkling ($1.00 in / $4.05 out) plus a 0.00-input PrismML model to its serverless card.
Detailsv0 by Vercel renamed its $30/user/month Team plan to Plus across the pricing page, docs and enterprise page. Prices and per-model token rates are unchanged.
DetailsWeights & Biases raised its cheapest Serverless Inference model, DeepSeek V4-Flash, from $0.01/$0.01 to $0.14 input / $0.28 output per 1M tokens, adding a cached tier.
DetailsxAI restructured Grok API token pricing into short-context and long-context columns. Crossing a model's long-context threshold reprices every token in that request at roughly 2x.
DetailsxAI lowered cached input on its grok-4.5 flagship from $0.50 to $0.30 per 1M tokens, a 40% cut that widens the prompt-caching discount on the newest Grok model.
DetailsZendesk repriced its Contact Center add-on from US$50 to US$83 per agent/month paid yearly and split it onto its own plan surface with a $33 Minutes Block and a $35 Quality Assurance add-on.
DetailsKai-Fu Lee's 01.AI is targeting a Hong Kong IPO in 2027 and repositioning toward enterprise/government deployment rather than open-model distribution.
DetailsTempus is acquiring molecular-residual-disease testing company Personalis for roughly $1.5B, folding MRD testing into its precision-oncology platform.
DetailsMistral AI reportedly acquired Emmi AI, extending it into industrial engineering and simulation.
DetailsMoonshot AI is seeking investor approval to begin an IPO process, targeting a Hong Kong listing within six months following the Kimi K3 model launch.
DetailsApollo rewrote its API docs with an explicit per-endpoint credit table, exposing a +8-credit mobile-phone charge and waterfall enrichment bands up to 45+ credits.
DetailsCoreWeave launched a $2.6B term loan to finance GPU purchases — continued debt-funded capacity expansion.
DetailsNebius raised $775M in its first secured debt financing — non-dilutive capital to accelerate its AI-cloud buildout.
DetailsFireworks AI raised a $1.5B Series D at a $17.5B valuation to expand enterprise AI inference infrastructure.
DetailsFlexprice raised a reported $1.5M seed round as AI companies rethink metering and billing.
DetailsLegal-AI company Harvey acquired Benchmark, an AI platform aimed at investment firms — its third acquisition in seven months, extending it beyond law firms into the buy side.
DetailsVercel acquired Egyptian-founded DevOps agent startup Stakpak — its second African acquisition of the month — to extend its agentic infrastructure line.
DetailsMoonshot shipped flagship Kimi K3 (2.8T parameters, 1M-token context) at $3.00 input / $15.00 output per 1M tokens, plus a Kimi K2.7 Code SKU at $0.95 / $4.00 with a double-priced HighSpeed variant.
DetailsSierra launched Horizon, replacing its Agent Data Platform line. Agents now pursue one business outcome across days or months, billed on results rather than tokens.
DetailsAnthropic moved closer to a large IPO, with bankers arranging investor meetings; reported 3Q26 profit exceeded $1B.
DetailsShield AI completed the acquisition of Aechelon Technology, a simulation and visual-systems supplier.
DetailsSuno raised $400M at a $5.4B valuation and says it is building toward IPO readiness, even as major-label copyright litigation intensifies.
DetailsDeepSeek is preparing an IPO filing and in talks to raise fresh capital at a ~$74B valuation, with annualized revenue reported near $500M.
Details6sense pulled the free Sales Intelligence tier (50 data credits/mo, Start for Free) from its pricing page; the page now shows only three quote-based Sales Intelligence bundles, all routing to Book a Demo.
DetailsArtisan removed its public credit-pool rate card for the second time, replacing Free/Intern ($250)/Employee ($600) with three quote-only tiers — Team (~2,500 leads/mo), Scale (~6,000), Enterprise — all 'Talk to sales.'
DetailsBright Data restructured Web Unlocker, SERP API, and Web Scraper API from a four-tier committed ladder to a Free / PAYG / Scale / Enterprise model, adding a recurring 5,000-results/month free tier (no card) and dropping the old $1.1 and $1.0 per-1k committed rates.
DetailsBright Data raised its Bright Insights eCommerce-intelligence list prices sharply: the eCommerce Tracker moved from $250/mo and Sales & Market Share from $400/mo to a $2,000/mo starting price for each — a 5x-8x increase on the published get-a-quote floor.
DetailsCovariant.ai now shows only a logo and copyright line; the marketing site and Contact-us sales motion are gone, signaling a wind-down after the 2024 Amazon deal.
DetailsDaily's Video SDK rate card added a dedicated PSTN phone-number SKU at $2 per number per month, and the discounted audio-only participant-minute rate now applies automatically instead of requiring an email to sales.
DetailsDeepInfra lifted dedicated GPU rates 16-32% (H100 to $2.20, B200 to $3.69, B300 to $4.89/GPU-hr) and raised DeepSeek-V3.1 to $0.25/$0.95 per 1M tokens, a notable reversal for a vendor known for cutting prices.
DetailsDify rebuilt dify.ai/pricing into one unified page: the old Cloud-vs-Self-hosted toggle is gone, Dify Cloud now sits beside a Self-Hosted & Enterprise column, and the Monthly/Annually switch defaults to Annually — exposing annual list prices (Professional $590/workspace/yr, Team $1590/yr; 17 pct off) that were previously gated. Monthly $59/$159 unchanged; new per-tier quota dimensions added.
DetailsExa raised its free tier from 1,000 to 20,000 requests per month and repriced the Agent endpoint: headline range now $0.012–$1.00/run, X-high cut from $2.00 to $1.00, plus a new $0.012 Minimal mode.
DetailsGoogle added AlphaEvolve, the DeepMind evolutionary coding agent, to Vertex AI pricing, billed as the Gemini model rate plus a 2x agent surcharge (3x all-in), e.g. 6 USD / 36 USD per 1M on Gemini 3.1 Pro Preview.
DetailsGorgias repackaged its AI Agent from a pay-as-you-go add-on into a bundled plan component with an included automated-interaction allotment, and raised the overage rate past the allotment from $0.90 to $1.50 per interaction.
DetailsGroq launched Orpheus text-to-speech billed per 1M characters ($22 English / $40 Arabic), a Browser Automation tool at $0.08/hour, and new models including GPT OSS Safeguard 20B, Qwen 3.6 27B, and Kimi K2.
DetailsLinkup adds a closed-beta Extract endpoint (structured crawl, variable ~$2-$10/task, $10 minimum balance) and corrects the pricing page Deep-search range from $0.05-$0.55 down to $0.05-$0.055, matching its billing docs.
DetailsMem added a third pricing tier, Mem Proactive, at $99/month (7-day free trial), bundling a new always-on Mem Agent. The old custom-quoted Mem Teams tier was dropped from the pricing page.
DetailsMercor Experts hero average contracted rate jumped from $80/hr to $122/hr in two weeks, and its Enterprise page now advertises a $2B+ revenue run rate, double the $1B+ shown through June 2026.
DetailsMetronome's free-to-start Starter plan now shows explicit rates: $100,000 billing volume and 10M events included per month, then 0.8% of billing volume and $0.04 per 1,000 events.
DetailsModal added the B300 (Blackwell Ultra) GPU at $0.001972/sec above the B200, plus a separately-metered Sandbox + Notebooks tier at ~3x standard CPU/memory rates, and now itemizes region (1.5-1.75x) and non-preemptible (3x) modifiers.
DetailsNetlify's Pro plan added a Choose a Pro tier selector: the $20/month base still includes 3,000 credits, with new selectable 5,000, 10,000, 15,000 and 20,000-credit monthly tiers above it. Headline plan prices unchanged.
DetailsOpenRouter's bring-your-own-key free allotment shifts from 1M requests/month (5M Enterprise) to $25,000/month of list-price inference ($200,000 Enterprise), keeping the 5% fee after.
DetailsPowerdrill re-packaged its Bloom credit model: every tier now includes 1,000 daily refreshed credits, and each paid tier reads 1,000 daily + N monthly credits (Pro 5,000 / Plus 11,000 / Premium 60,000). Headline monthly prices were unchanged.
DetailsResemble AI stripped all voice-generation per-second rates and voice-clone add-ons from its pricing page, leaving only deepfake-detection, verification and watermarking usage; it also dropped the published 80% Enterprise discount and raised the sales threshold from 500 to 2,000 dollars per month.
DetailsRunPod trimmed two Secure Cloud Pod rates: H100 SXM (80GB) dropped from $3.29 to $2.99/hr and RTX Pro 6000 (96GB) from $2.09 to $1.99/hr, while the rest of the per-hour, Serverless, Clusters, and storage rate cards held.
DetailsSmartlead lifted every plan - Base $32 to $39, Pro $78 to $94, Smart $144 to $174, Prime $315 to $379 - while cutting Prime's send cap from 5.69M to 500K/mo and free verified emails from 2.04M to 170K.
DetailsTogether AI launched a Provisioned Throughput (PTU) SKU at $0.05/PTU-min and cut Dedicated Inference rates — on-demand H100 to $5.49/hr, B200 to $8.99/hr.
Detailsturbopuffer reduced the entry-tier Launch plan monthly minimum from $64 to $16, a 75% cut to its cheapest paid floor, for all new and existing customers.
DetailsUnstructured restructured its per-page Serverless API: the 15,000 free pages now reset every month instead of being a one-time allowance, and Pay-As-You-Go bills are capped at $3,000/mo — above which pages are free up to 1 million pages a month.
DetailsWeights & Biases added ARIA — a token-based-pricing product — as its own metered line on the pricing compare table, free for a limited time on Free and Pro and included on Enterprise. Core tiers and meters unchanged.
DetailsxAI shipped grok-4.5, an intelligence-first flagship (500k context) priced at $2 in / $6 out per 1M tokens — its first new flagship priced above the prior one, with grok-4.3 kept live at $1.25 / $2.50.
DetailsOxylabs raised $130M from Warburg Pincus at a $3.6B valuation.
DetailsHootsuite now frames its plans as a Social OS built from Wisdom AI, Perch, Nest, and Lumen, bundling an in-app AI teammate into every tier from Standard to Enterprise rather than selling AI as a metered add-on.
DetailsQdrant raised a $50M Series B to position composable vector search as core production-AI infrastructure.
DetailsBeijing intervened to block Meta's ~$2B acquisition of Manus, and Tencent is leading an effort to unwind it and acquire the agent startup instead. Ownership is unresolved — a material continuity risk rather than a pricing event.
DetailsMercor acquired a16z-backed Deeptune to build training environments/simulators for AI agents, extending it from talent marketplace into RL environment supply.
DetailsMiniMax raised $2B via combined share and bond sales, funding continued open-weight model development.
DetailsOpenAI's deployment arm is acquiring Northslope, per an Axios exclusive, while IPO-timing coverage suggests the listing window may slip to 2027.
DetailsSambaNova raised a $1B Series F led by General Atlantic at an $11B valuation, five months after its previous mega-round, and named JPMorgan as an infrastructure customer.
DetailsSuperhuman launched Docs, extending it from email into AI-assisted document collaboration — a scope expansion on an unchanged per-seat price.
DetailsZhipu AI raised $4B via a Hong Kong share offering months after its January IPO, following a reported ~1,500% stock rise.
DetailsPortkey, the LLM gateway/observability layer, is being acquired by Palo Alto Networks. Deal terms were not disclosed in the trade coverage; the standalone gateway pricing page showed no change in this run.
DetailsxAI completed a rebrand to SpaceXAI with a new logo as it fully merges into SpaceX. Separately, Musk acquired gas-turbine operator APR Energy (~$1B) to power the training fleet.
DetailsAnthropic launched Claude Sonnet 5 as the new production-standard Sonnet tier at introductory $2/$10 per 1M tokens through Aug 31, 2026, then standard $3/$15 — undercutting the long-held Sonnet band during the intro window.
DetailsAssemblyAI added a Voice Agent API at $4.50/hr ($0.075/min), broke Guardrails out as a priced family, and reorganized its pricing page into six product tabs while keeping async rates flat.
DetailsFlexprice's cloud Free plan now advertises up to $100K cumulative billing revenue in place of its prior 1-month validity, and the $1000/mo Scale tier switched its CTA from self-serve Get Started to Contact Us.
DetailsGoogle's consumer AI plan lineup now includes Google AI Plus (400 GB storage, 2x usage limits, Pro-model access), a new entry paid tier sitting below the $19.99 AI Pro plan.
DetailsDynamics 365 keeps its per-user per-app seat ladder ($50–$300/mo) but now meters autonomous agents separately in Copilot Credits — 1 credit per classic answer, 2 generative, 5 per agent action, 10 for tenant-graph grounding.
DetailsMistral increased developer-API prices while holding Vibe consumer plans flat: Mistral Small 4 rose to $0.15 in / $0.6 out per M tokens and the OCR model (now OCR 4) doubled to $4 per 1,000 pages ($5 with Document AI).
DetailsOxylabs dropped the Web Scraper API Regular/Enterprise tab split for a single Micro/Starter/Business/Custom+ grid, removing the Advanced, Venture and Corporate named tiers from the page. Per-1K rates on surviving tiers are unchanged.
DetailsPipedrive now sells four per-seat CRM plans - Lite $14, Growth $24, Premium $49 and Ultimate $69/seat/mo (annual) - each marketed as Now with AI, replacing its older five-tier lineup.
DetailsRunPod repriced Serverless flex GPU workers — H200 up to $5.93/hr, H100 to $4.55/hr, RTX 6000 Pro cut to $3.49/hr, L40-class to $1.75/hr — and added B300 ($9.98/hr) and RTX PRO 4500 Blackwell ($1.15/hr) worker SKUs.
DetailsSnowflake moved Cortex AI features onto a new AI Credit meter priced by routing ($2.00 global / $2.20 regional), decoupling AI cost from Snowflake edition; Platform Credits ($2/$3/$4) still cover warehouse compute and storage.
DetailsSpeechmatics cut real-time enhanced speech-to-text from $0.56 to $0.43/hr, added a $0.129/hr multilingual Batch Melia 1 model, and grew the free tier to 3,000 minutes (50 hours) per month.
DetailsVercel Agent appears on the main pricing page as a BETA AI SKU billed $0.25 per million tokens plus pass-through model cost on Pro, with Code Review and Investigations capabilities.
DetailsWeaviate removed its Plus plan ($280/mo, SSO, 99.9% uptime); Weaviate Cloud now ships three tiers — Free, Flex ($45/mo PAYG) and Premium ($400/mo prepaid) — with SSO, 99.95% uptime and shared-or-dedicated deployment consolidated into Premium.
DetailsHubSpot acquired Warmly, a signal-based buyer-intent tool, to close a CRM visibility gap.
DetailsTogether AI raised $800M at an $8.3B valuation, with the CEO declining to rule out an IPO next year. The raise underwrites continued GPU price aggression.
DetailsSquarespace narrowed its website plans to Basic ($19), Core ($29) and Advanced ($99) — dropping the former Plus tier — with Blueprint AI and Squarespace AI bundled across the lineup.
DetailsAgility Robotics is listing on Nasdaq through a $2.5B SPAC merger with Churchill Capital.
DetailsCreatify raised its Platform plans — Starter $33 to $39/mo and Pro $49 to $99/mo — while cutting the annual-billing discount from up to 50% to up to 15% and capping the Pro credit-pool slider at 1,000 credits.
DetailsDeepInfra introduced a per-request Service Tier on its pricing page: Standard stays 1× base price, while a new Priority tier bills 1.5× for faster time-to-first-token during peak demand.
DetailsHume AI exposed a model toggle on its pricing page: EVI 4 MINI runs at roughly half the EVI 3 per-minute overage and Octave 2 roughly halves Octave 1 per-character overage while doubling minutes per character. Tier prices are unchanged.
DetailsIntercom's June-2026 new-customer offer cutting Essential 35% ($19/seat, $228/yr) has ended; the pricing page and calculator now show Essential at its flat $29/seat list rate ($348/yr). Advanced, Expert, and the $0.99/Fin meter are unchanged.
Detailslemlist's flat-fee Email plan (unlimited users) jumped from $39 to $69/mo; its included email allowance rose from 5,000 to 50,000/mo. Multichannel held at $109/user/mo.
DetailsLokalise dropped its Growth plan from $499 to $375/mo (annual billing) and removed the $0 Free tier, making Explorer ($144/mo) the entry plan; Growth's bundled allowances also shrank.
DetailsMercor renamed its Experts hero metric from average pay to average contracted rate and dropped the figure from $141/hr to $80/hr, while roles created climbed to 287.1K and buyer-side pricing stayed fully sales-gated.
DetailsQodo repackaged its code-review platform onto a team-wide credit pool at $.012/credit with selectable packs, retiring per-seat Teams pricing and the permanent free Developer tier in favor of a 14-day trial.
DetailsRecraft gave its newest V4.1 models a separate, lower API tier: V4.1 raster dropped to $0.035 (from $0.04) and V4.1 Pro raster to $0.21 (from $0.25).
DetailsRunPod added the Blackwell-Ultra B300 (288GB HBM3e) at $7.39/hr as the new top of its Secure Cloud Pod ladder, above B200 ($5.89/hr), and re-sorted L40S to $0.99 and L40 to $0.82.
DetailsShortwave repriced all four paid seats up — Pro $14 to $18, Business $24 to $30, Premier $36 to $45, Max $100 to $120 per seat/mo — and dropped the annual toggle so the page now quotes a single monthly per-seat price.
DetailsTogether cut GPU Cluster rates: on-demand HGX H100 to $3.99/hr (from $4.79) and reserved 7–30 day H100 to $3.59/hr (from $4.19), with the 91–180 day reserved floor at $3.09/hr.
DetailsVercel's v0 Max Fast model drops to $10/1M input and $50/1M output tokens (from $30/$150) and the legacy Ultra/Premium tier is retired, leaving Free, Team ($30), Business ($100) and Enterprise.
DetailsGitHub lifted the April 2026 pause on new individual Copilot sign-ups. The plans page again shows a self-serve Get started button on Free, Pro, Pro+ and Max.
DetailsHumanoid-robotics maker Agility Robotics announced it will go public through a merger with Churchill Capital Corp XI, filing a confidential Form S-4 draft registration in July 2026. Its Digit RaaS pricing stays sales-only.
Detailscontinue.dev/pricing now serves an acquisition announcement that Continue has been acquired by Cursor, with no plan pricing. The hosted Starter, Team and Company tiers are no longer published; the open-source extension stays free.
DetailsDust scrapped its flat €29/seat unlimited-message plan for a USD, credit-metered model with three seat types — Free ($0), Pro ($30/mo, $24 yearly), Max ($150/mo, $120 yearly) — plus its first free tier.
DetailsFrase held its $49/$129/$299 prices but added per-plan meters (AI generations, Content Guard pages, API caps), productized Frase Answers, and switched overage to optional pay-as-you-go that is off by default.
DetailsIntercom now runs a new-customer offer taking the Essential plan 35% off, from $29 to $19 per seat/month ($348 → $228/yr). Advanced ($85), Expert ($132), and the $0.99/Fin-outcome meter are unchanged.
DetailsNovita on-demand RTX 4090 GPU instances drop to $0.35/hr ($0.18 spot) from $0.67/$0.34; L40S and H100 instances removed, H100 moves to dedicated/bare-metal only.
DetailsRunway retired its $95 Unlimited tier and replaced it with Max at the same price, swapping unmetered Explore-mode generation for a 9,500-credit monthly allowance (up from 2,250) plus 1-month credit roll-over.
DetailsSynthflow pulled its self-serve per-minute pay-as-you-go path off the pricing page and now shows a single Enterprise plan — contracts from $30,000/year, scoped and quoted via Contact Sales.
DetailsTavus restructured its developer Conversational Video Interface plans into six tiers — adding a $22/mo Starter entry plan and a $975/mo Business tier, renaming the old $59 plan to Builder, and cutting per-minute overage rates.
DetailsTogether cut GPU cluster rates (on-demand H100 $5.49→$4.79, reserved 7–30d H100 $4.99→$4.19), repriced serverless models, and began publishing cached-input rates on serverless inference.
DetailsWellSaid Labs replaced its single $50 Creative seat with two new tiers — Starter $10/mo and Pro $33/mo (annual) — and switched its capacity meter from downloads/year to minutes/year of downloaded audio.
DetailsZapier converged its meter: AI steps, code and the SDK now follow the same task-based model and share one task pool with Zaps and MCP, with task rates varying by AI model tier, code runtime and connector type.
DetailsAiSDR moved from credit-and-message tiers to AI-researched-contacts metering (Solo $250 / Explore $900 / Scale $2,500) and added a new $250 month-to-month entry plan.
DetailsCanva now excludes Leonardo.Ai Essential features during a Canva Business free trial; the bundled access enables only once the trial converts to a paid subscription.
DetailsDocket prices its AI marketing agent by monthly website visitors: Growth from $3K/mo, Scale from $4K/mo, Enterprise custom — all-inclusive, with no seat fees or conversation caps.
DetailsLangfuse's Stripe case study discloses that it started with an internal billing system based on a simple per-event flat rate, then migrated to Stripe Billing (usage-based), Stripe Checkout, and Stripe Tax for the rails. It keeps the metering in-house: it posts hourly event counts from its own ClickHouse/OLAP data store to Stripe's metered-usage API, handling up to 200M events/month per account — the operational mirror of its "unit = trace + observation + score" price metric. A "Senior Backend Engineer (IAM and Billing)" req (open in the 2026-04-17 ATS snapshot, since closed) staffed that billing platform; growth analytics run on a bought PostHog/BigQuery/dbt/Metabase stack.
DetailsPinecone's original billing system was built in-house but didn't support multi-product billing or complex metering — the team still calculated some invoices manually. When it launched serverless (read units / write units / storage), Pinecone adopted Orb as the system of record for all usage and billing data, enabling add-ons, discounts, custom plans, and version-controlled pricing without engineering bottlenecks. The RU/WU metering layer remains in-house; Orb handles invoicing and plan management.
DetailsAnthropic job posts disclose a homegrown ledger application as the center of its finance systems, with a Finance Systems Engineer req stating the team is "moving beyond configuring off-the-shelf platforms and into building homegrown, production-grade financial applications that no vendor has built for us yet." It still buys the surrounding stack — Salesforce + Salesforce CPQ (CRM/quote-to-cash), Zuora and Stripe (billing/payments), Metronome (usage metering), NetSuite (ASC 606 rev-rec), and Tesorio (AR automation) — and is staffing a dedicated Billing Platform engineering team (~9 open billing-eng roles).
DetailsOpen roles — a dedicated Billing engineer building the usage-metering pipeline (raw events → invoiced amounts in Stripe), a Senior Staff Finance Systems engineer, and ~10 RevOps roles — show Cursor building the meter behind its seat-plus-usage pricing while buying payments.
DetailsIts 'Beyond rate limits' engineering post describes a home-grown engine fusing rate limits, real-time usage tracking, and prepaid credit balances — the spend-control layer most companies buy — alongside a dedicated Financial Engineering hiring push.
DetailsAnthropic launched Claude Fable 5 (GA June 9, 2026) and Claude Mythos 5 (limited availability via Project Glasswing) at $10/$50 per 1M tokens — a new top API band above Opus 4.8, double its $5/$25 rate, with a 1M context window.
DetailsAnthropic added a research-preview Fast mode premium-speed SKU on the Opus tier (Opus 4.8 $10/$50; Opus 4.6/4.7 $30/$150) and Claude Managed Agents, which bills $0.08 per session-hour of runtime as a second meter on top of standard token rates.
DetailsIdeogram's API now lists an Ideogram 4.0 model tier ($0.03 Turbo / $0.06 Default / $0.10 Quality per image, plus 4.0 Custom inference at $0.06–$0.20), alongside new Generate-with-Gemini, Instructional Edit, Topaz Upscale and Layerize endpoints. The consumer subscription ladder (Free/Plus $15/Pro $42/Team $20-seat/Enterprise) is unchanged.
DetailsSalesforce signed a definitive agreement to acquire Fin — the company that renamed from Intercom to Fin in May 2026 — for approximately $3.6 billion. Expected to close in Q4 of Salesforce's FY2027 (~early 2027), subject to regulatory approval. Fin's AI Agent and proprietary Apex support model will complement Salesforce Agentforce; customer-facing pricing is unchanged at announcement (Fin $0.99/resolution; Essential $29 list).
DetailsAdyen announced a definitive agreement on 2026-06-11 to acquire usage-based billing platform Orb for $335M all-cash (reverse triangular merger), closing in early July 2026. Unlike a wind-down, this is a continuity event: Adyen is running an 'incubator' model that preserves Orb's operations and multi-PSP support in the first phase, so Orb's three custom-priced tiers (Core / Advanced / Enterprise, billed on billings + events + platform fee) are unchanged for now. The strategic aim is to link usage-based billing data directly to payment processing.
DetailsLinear launched Coding Sessions (June 11, 2026), its first usage-metered feature — billed from a workspace-level prepaid AI-credit balance shared across AI tools.
DetailsGitLab announced Flex: one annual dollar commitment covering platform seats and GitLab Credits, reshaped month-to-month without contract amendments.
DetailsClose brought its Chloe AI sales agent to general availability, included at no extra SKU on all four CRM tiers and metered instead through per-user AI credits plus telephony usage.
Detailssourcegraph.com/pricing now shows a single Enterprise plan starting at $16K with seat-scaled, org-wide pooled AI credits. The self-serve Cody consumer tiers are gone, and the standalone agentic successor Amp was spun out to ampcode.com with separate billing.
DetailsPhind shut down its AI developer search service on 2026-01-16 (data deleted 2026-01-30, prorated Pro refunds) — roughly six weeks after a ~$10.4M raise (Bessemer + YC, reported 2025-12-03). As of 2026-06-08 phind.com/plans returns a Vercel 404 DEPLOYMENT_NOT_FOUND. Final plan prices are third-party indicative (Free $0; Pro ~$20/mo; Business ~$40/user/mo) because the app rendered prices from a runtime API web archives never captured.
DetailsClipdrop's developer API pricing page (clipdrop.co/apis/pricing) no longer publishes per-call or credit rates. It now renders a Jasper Contact-us form to request additional credits or migrate to Jasper's API services, reflecting Jasper's ownership of the product. Consumer Free and flat-fee Pro tiers remain self-serve on clipdrop.co/pricing.
DetailsOxylabs is discounting Web Unblocker by 40% for six months via coupon code WU40 at checkout, cutting the entry Micro plan from $9.40/GB ($75/mo) to $5.64/GB ($45/mo) and applying across all Regular and Enterprise Web Unblocker tiers.
DetailsScraperAPI acquired Traject Data, bringing dedicated SERP and e-commerce data APIs into its product line and extending the same API-credit metering across structured search-results and marketplace data.
DetailsWellSaid Labs moved its canonical domain and pricing surface from wellsaidlabs.com to wellsaid.io (the old domain 301-redirects). The live plan lineup on /ai-voice-pricing is Trial (free 7-day) / Creative ($50/mo/user annual, $55 monthly) / Business ($160/mo/user) / Enterprise (custom), each with a fixed annual download quota.
DetailsSalesforce agreed to acquire composable-content platform Contentful (reported ~$1–1.5B), folding its usage-metered headless CMS and pay-as-you-go AI Actions into Salesforce's ecosystem.
DetailsMurf renames its Studio plans to Creator and Business and ships the Murf API and Murf Dub as separate usage-billed products.
DetailsYou.com raises Research Lite to $12 per 1k and ships a dedicated Finance Research API, extending its 2026 API repricing.
DetailsIdeogram sets a four-tier subscription grid plus a per-image pay-as-you-go API from $0.025.
DetailsSynthesia lowers prices across its lineup, with self-serve plans now starting at $14/mo.
DetailsAnthropic refreshes Claude.ai with the Claude 4.x family and adds a $200/mo Max tier above Pro.
DetailsCerebras introduces fixed-price Cerebras Code coding subscriptions alongside a tightened public inference rate card.
DetailsHeyGen moves every plan onto a single credit currency spanning avatars, dubbing and video.
DetailsSemrush replaced its long-running Pro/Guru/Business tiers with a four-tier 'SEO + AI Search' lineup (Free/SEO/Starter/Pro+/Advanced), bundling AI-search-visibility tracking into every tier.
DetailsWebflow collapsed its Site plans to three tiers (Starter/Basic $15/Premium $25), added a $2,500/mo Team Platform Workspace plan, and bundled AI credits into every Workspace with caps enforced from June 2026.
DetailsCharacter.ai removes legacy chat styles for free users and soft-launches them behind its c.ai+ subscription.
DetailsElevenLabs cuts Conversational AI pricing again as it leans further into pay-as-you-go agent usage.
DetailsHeyGen renames its core credit and makes audio dubbing unlimited ahead of the credit-model cutover.
DetailsDocuSign increased list prices across its self-serve eSignature and IAM web plans: eSignature Standard rose to $30/user/mo and Business Pro to $45/user/mo, and IAM Professional to $80/user/mo. Effective date approximate (bracketed to Q2 2026; Wayback was unreachable to pin the exact date).
DetailsMistral renames Le Chat to Vibe, repositioning it as a work-and-code agent, and introduces a $5.99/mo student plan.
DetailsOpenAI ships the GPT-5 family — GPT-5.5 at $5/$30 and GPT-5.4 at $2.50/$15 per 1M tokens — its biggest API reset since GPT-4o.
DetailsWith its Spring 2026 Omni release, Outreach rebranded from outreach.io to outreach.ai and shifted its gated packaging to Amplify Core/Plus/Pro, layering consumption-based AI credits (25k/50k/100k) on top of the per-seat license.
DetailsGitHub paused new self-serve purchases of Copilot Business ($19/user/mo) and Enterprise ($39/user/mo) on April 22, 2026; new org customers must now contact sales while self-serve is said to be reopening soon.
DetailsNetlify removed per-seat charges from its credit-based Pro plan: previously $20 per seat (including active Git contributors), now a flat $20/month for 3,000 credits with unlimited team-member seats. CEO Mathias Biilmann framed the move around AI development that doubled Netlify community in one year past 8 million developers.
DetailsSugarCRM announced its new brand identity as SugarAI on 2026-04-13 (BusinessWire) — its first name change since founding in 2004 — repositioning around AI-driven 'precision selling.' The corporate and pricing domains now redirect to sugarai.com and the pricing page foregrounds predictive/generative AI. Crucially, the rebrand did not change the money: Sugar Sell stays at Standard $59 / Advanced $85 / Premier $135 per user/mo (billed annually, 15-user minimum), and AI (SugarPredict / Sugar Intelligence / sales-i) remains bundled into the Advanced and Premier editions rather than metered as usage or sold as credits. Market, Serve, Enterprise and the sales-i add-on stay contact-sales.
DetailsSprout Social introduced a new publishing-focused Essentials plan at $79/seat/month on annual billing ($99/seat/month billed monthly), positioned below the existing Standard tier. Essentials includes up to 5 social profiles, optimal send times, and profile/post-level reporting, but excludes the consolidated inbox and is not eligible for add-on modules. The change dropped Sprout's advertised starting price from $199 to $79. The tier was absent from all Wayback snapshots through 2026-03-13 and present by 2026-04-10.
DetailsCisco announced its intent to acquire AI-observability and evals platform Galileo on 2026-04-09 to extend Splunk Observability Cloud; the deal closed 2026-05-22 and galileo.ai now carries a now part of CISCO lockup. Terms undisclosed.
DetailsExa restructures pricing into per-endpoint cards and raises base Search to $7 per 1,000 results.
DetailsGoogle releases the Gemini 3 model family and introduces region-based token pricing.
DetailsWindsurf (now Cognition) replaced its prompt-credit system with a usage allowance that refreshes daily and weekly, raised Pro from $15 to $20/mo, added a $200/mo Max tier, and moved overage to raw model API pricing. Wayback snapshots show credits live on 2026-03-08 and the quota model live by 2026-04-01; the no-rollover change drew a 33% price-hike backlash on Reddit. The surface has since folded into Devin Desktop, with windsurf.com/pricing 308-redirecting to devin.ai/pricing following Cognition's July 2025 acquisition.
DetailsYou.com slashes API prices, dropping Search to $5/1k and Contents to $1/1k, and retitles its pricing page.
DetailsPoe's Free tier's daily compute-points allowance fell from ~3,000/day to ~300/day around 2026-03-30, a ~90% cut Poe never announced. The date and magnitude are corroborated by three independent third-party sources, not a vendor confirmation.
DetailsSalesforce hired the Clockwise team into its Agentforce division (announced 2026-03-20), explicitly stating it was not acquiring the product or technology. The AI calendar-optimization product becomes unavailable on 2026-03-27; prepaid subscriptions are refunded prorated and all user data is deleted with no transfer to Salesforce. Clockwise endorses Reclaim.ai (100% price-match guarantee) as the migration path.
DetailsGumloop changed credit overage from a hard cap to unlimited-by-default (Enterprise can set a custom cap) and merged its Solo and Team plans into one seat-free Pro, days around its $50M Series B.
DetailsPerplexity makes its Comet browser free and introduces a $5/mo Comet Plus add-on.
DetailsUpstash Box adds a sixth product line: sandboxed computers for AI agents priced at $0.10 per active CPU hour, Upstash's first compute-metered SKU alongside its five data products.
DetailsFlow Studio's Wonder 3D adds text-to-3D and image-to-3D asset generation billed at 20 credits each out of the existing monthly credit pool — no new SKU, no rate-card change.
Detailsm3ter announced a strategic partnership with and investment from Salesforce on 2026-03-04 (BusinessWire / m3ter blog, London dateline). Salesforce selected m3ter as an advanced metering and rating partner for Revenue Cloud Advanced (RCA), Revenue Cloud Billing (RCB), Agentforce Sales, and CPQ, so Agentforce Revenue Management customers can run high-scale rating, mediation, and usage-pricing models. The investment amount and round were not disclosed. This is NOT an acquisition: m3ter remains an independent company (M3ter Holdings Limited), keeping its own brand, roadmap, and gated four-component pricing.
DetailsModal adds AWS and GCP Marketplace billing so enterprises can pay through existing cloud commitments.
DetailsSuperhuman (the renamed Grammarly) acquired AI-spreadsheet company Rows on 2026-02-22. Unlike a price-preserving acquisition, this is a product sunset: rows.com fully winds down on 2026-05-31 after 9 years and 2.2M users, the team folds into Coda, and the pricing page now routes users to 'Continue with Coda.' Superhuman Terms apply from 2026-06-16.
DetailsBaseten adds discounted cached-input pricing to its multi-tenant Model APIs.
DetailsA Lovable Financial Systems Engineer req discloses a new in-house Financial Engineering team owning 'the billing system of record—from Stripe integration through to the metering ledger' and 'a credit-based usage model, pass-through service billing.' Signal: Lovable keeps Stripe for payments but builds its complexity-weighted credit meter in-house rather than buying a usage-billing platform (no Metronome/Orb/Chargebee named).
DetailsRunPod begins charging different rates for idle versus running network-volume storage.
DetailsNebius agreed to acquire AI-search-API startup Tavily for a reported ~$275M (up to $400M on milestones). Tavily's public credit-based API pricing (Researcher free, $0.008/credit pay-as-you-go, $30/mo Project) was unchanged through the announcement — the deal is a distribution/infra bet, not a repricing. Founder Rotem Weiss joins Nebius.
DetailsCartesia takes Voice Agents to general availability on a flat per-minute rate.
DetailsFireworks introduces parameter-size-based pricing for its embeddings models.
DetailsA Synthesia engineering post details a billing-and-usage-metering system the team built in-house in ~3 months and ran on roughly one engineer while scaling from $40M to ~$140M ARR, including the 2025 move to a unified credit-based model. Stripe is kept strictly as the payment provider behind an isolated data model; Salesforce remains the CRM. The metering layer is built, not bought.
DetailsCursor settles into a six-plan lineup following its 2025 credit-pool transition.
DetailsRecraft ships V4 and adds video generation across its credit tiers.
DetailsGroq adds per-use pricing for built-in agent tools — web search, page visits and code execution.
DetailsPerplexity splits its enterprise offering into Enterprise Pro and Enterprise Max tiers.
DetailsTogether AI adds a dedicated fine-tuning tier for specialized models.
DetailsStripe completed its acquisition of usage-based billing infrastructure company Metronome on 2026-01-14 (announced 2025-12-02), positioning Metronome as Stripe's metering-and-monetization-infrastructure layer. Deal terms were not officially disclosed; trade press reported ~$1B (unconfirmed). Metronome continues to operate its own free Starter and sales-quoted Custom plans.
DetailsCodeium lowers Windsurf Pro to $15/mo and expands the agentic credit pool after the OpenAI acquisition.
DetailsPerplexity reorganizes its developer API around a Search API and a new Agentic Research tier.
DetailsGoogle filed suit against SerpApi over scraping of its search results — a notable trust/regulatory event for the search-API category. SerpApi bundles a U.S. Legal Shield (up to $2M coverage) on Production+ plans, a differentiator the suit puts in the spotlight.
DetailsElastic (NYSE: ESTC) completed its acquisition of Jina AI; founder Han Xiao became Elastic VP of AI. Jina's standalone search-foundation APIs (Embeddings, Reranker, Reader, DeepSearch) and their token-credit pricing continue to operate independently post-deal.
DetailsGalileo repackaged pricing from two tiers (Developer $0 with up to 3 users / custom Enterprise) to three (Free $0, Pro $100/mo for 50,000 traces, custom Enterprise). The free tier moved from a 3-user cap to unlimited users; the change is visible between the 2025-09 and 2025-10 Wayback snapshots of galileo.ai/pricing.
DetailsCoreWeave (NASDAQ: CRWV) announced a definitive agreement to acquire fine-tuning/RL platform OpenPipe (terms undisclosed), citing its open-source ART (Agent Reinforcement Trainer) toolkit. OpenPipe's per-token training and inference rate card was unchanged post-acquisition.
DetailsOn 2025-09-03 Kong Inc. announced its acquisition of OpenMeter (YC W23); co-founders Peter Marton and András Tóth joined Kong. OpenMeter Cloud was rebranded to Kong (Konnect) Metering & Billing and, by early 2026, the openmeter.io/pricing page was reduced to a migration announcement with no plan grid or public prices. The free Apache-2.0 self-hosted edition remains. Sources: konghq.com/blog/news/kong-acquires-openmeter and openmeter.io/blog/openmeter-is-joining-kong (both 2025-09-03).
DetailsMake rebranded its core usage metric from 'operations' to 'credits'. For non-AI apps 1 operation still equals 1 credit, but AI apps now consume credits dynamically based on tokens, file size, pages, or run time. Make states existing plans and pricing remained unchanged.
DetailsWindsurf's ownership split three ways in one week: OpenAI's reported $3B acquisition collapsed on 2025-07-11, Google paid ~$2.4B to hire founders Varun Mohan and Douglas Chen into DeepMind the same day, and on 2025-07-14 Cognition (maker of Devin) acquired the remaining company — product, brand, IP, ~$82M ARR, 350+ enterprise customers and a ~250-person team. Pricing was unchanged short-term; Windsurf has since been folded into Devin Desktop.
DetailsTechCrunch reported that a16z- and Benchmark-backed 11x displayed logos of companies that said they were not customers (ZoomInfo and Airtable both denied being customers; ZoomInfo threatened legal action over the unauthorized logo) and counted full annual contract value as ARR even for customers who exercised a ~90-day break clause — so a stated ~$14M ARR may have reflected only ~$3M of contracts that cleared the trial. An employee cited 70-80% churn of inbound customers; 11x said retention was 79%. Founder-CEO Hasan Sukkar stepped down six weeks later (May 2025); CTO Prabhav Jain became CEO.
DetailsMongoDB acquired embeddings/reranker company Voyage AI for ~$220M. Voyage's standalone per-token API and 200M free-token allowance were preserved post-acquisition; MongoDB Atlas became an additional distribution path rather than a forced migration.
Details01.AI's public rate card dropped from seven priced SKUs to two — yi-lightning (¥0.99/1M tokens) and yi-vision-v2 (¥6/1M) — with no surviving rate changed. Both are now intelligently routed to DeepSeek and Qwen models behind a fixed price.
DetailsThryv Holdings (NASDAQ: THRY) announced the closing of its acquisition of Infusion Software, Inc. (Keap) on 2024-10-31 for $80M in cash, subject to adjustment (definitive agreement announced 2024-10-29). Keap carried roughly $85M of trailing-twelve-month revenue through June 2024. Keap continues to operate its SMB CRM + marketing-automation platform under its own brand out of Grapevine, TX; the pricing page now footers a Thryv acquisition link and shares Thryv legal terms.
DetailsHubSpot unified its AI features under the Breeze brand and introduced HubSpot Credits, a single consumption pool that meters Breeze Agents and other AI features separately from per-seat subscriptions.
DetailsSalesforce introduced Agentforce, its autonomous-agent platform, priced on consumption for AI-agent work rather than per human seat — reframing AI pricing as digital labor atop its seat-based CRM.
DetailsZuora announces a definitive agreement to acquire usage-based metering and rating startup Togai, extending its monetization suite; deal expected to close May 2024, terms undisclosed.
DetailsNo changes match these filters.