Ask

DeepInfra Pricing Calculator

Updated August 2026

Estimate DeepInfra costs for Llama 4 Maverick, Llama 4 Scout, DeepSeek V4-Pro, DeepSeek V4-Flash, and Qwen3.7-Max.

DeepInfra pricing: DeepInfra pricing 2026: per-token inference (DeepSeek-V3.1 $0.25/$0.95, Llama-3.1-8B $0.02/$0.04), B200 GPU $3.69/hr, DeepCluster from $1.98/GPU-hr. Use the free calculator below to enter your usage and get an instant all-in monthly estimate — including overages — and see which plan is cheapest for your needs.

How DeepInfra prices: Pure-usage per-token inference + per-hour GPU instances + reserved multi-year GPU clusters + startup credits.

Best fit at your usage: DeepSeek V4-Flash

$0.54/mo— saves $1.26/mo vs your selection

Token Usage

Average number of tokens you send per API call

toke
1001M

Average number of tokens the model generates per call

toke
100500K

Request Volume

How many API calls do you make daily?

requ
1100K

Compare Plans at Your Usage

Monthly Estimate

Llama 4 Maverick

$1.80

per month

Per request: $0.0006

Cost Breakdown

Input tokens
$0.60
Output tokens
$1.20

Daily

$0.06

Annual

$21.60

View DeepInfra pricing page

Why DeepSeek V4-Flash$0.54/mo

  • DeepSeek V4-Flash at $0.54/mo.
  • Input tokens is 50% of your bill — the lever that matters most.

Budget range

Plan for usage swings, not just today's estimate.

Conservative

$1.50

500 tokens/mo

Expected

$1.80

1K tokens/mo

Aggressive

$2.40

2K tokens/mo

Need a calculator like this on your pricing page?

Embed interactive pricing calculators on your website to help customers understand costs and boost conversions.

Get Started

About this DeepInfra calculator

This calculator estimates your DeepInfra cost from publicly available pricing. Actual costs may vary with your specific agreement, volume discounts, and usage patterns — always verify on the provider's official pricing page for the most current rates.

DeepInfra pricing — frequently asked questions

How much does DeepInfra cost per token?

Per-1M-token rates vary by model. Examples: DeepSeek-V3.1 $0.25 in / $0.95 out, DeepSeek-V4-Pro $1.30 / $2.60, Llama-3.3-70B-Turbo $0.10 / $0.32, Llama-3.1-8B $0.02 / $0.04, Gemini 2.5 Pro $1.25 / $10.00, Kimi-K3 $2.85 / $14.25. Rates span from $0.02 per 1M input tokens (Llama-3.1-8B) to $2.85 (Kimi-K3, $14.25 out), the most expensive row now on the /pricing page. As of 2026-08-28 the page's Claude section lists only claude-haiku-4-5 ($1.00 / $5.00) — six other Claude models, including the $10.00/$50.00 claude-fable-5 row that had been the single highest rate on the list, no longer appear (they were still listed as of 2026-08-26). Many models also show a discounted cached-prompt input rate inline. These are the default Standard-tier (1×) rates; a per-request service tier multiplies them — Flex bills 0.8× for non-production and asynchronous work, Priority bills 1.5× for faster time-to-first-token.

Does DeepInfra charge per GPU-hour for dedicated hardware?

Yes. Custom-LLM deployments on dedicated GPUs bill in minute granularity (invoiced weekly): A100 $0.89, H100 $2.20, H200 $2.69, B200 $3.69, B300 $4.89 per GPU-hour. The on-demand GPU-instances product lists 1×B200 $3.69/hr, 2×B200 $7.38/hr, 4×B200 $14.76/hr, 8×B200 $29.52/hr with no egress fees.

What is DeepInfra DeepCluster pricing?

DeepCluster is a reserved, customer-owned NVIDIA B300 GPU cluster (256–5,000 GPUs) that DeepInfra procures and operates. All-in pricing is $2.99/GPU-hr on a 3-year term and $1.98/GPU-hr on a 5-year term, marketed as up to 54%–70% cheaper than a $6.50/GPU-hr public-cloud reference. Terms are 3 to 5 years and are sales-led (contact [email protected]).

Does DeepInfra have a free tier?

No published free tier — a card on file or pre-payment is required before you can use the service. However, the DeepStart program grants qualifying startups 1,000,000,000 free tokens (valued at DeepSeek-V3.1 prices) for companies that raised $250K–$10M and were founded within the last 2 years.

Is this DeepInfra pricing calculator free and accurate?

Yes — it's completely free with no signup. It uses the latest publicly available DeepInfra pricing (verified August 2026). Actual costs may vary with volume discounts or enterprise terms — always confirm on DeepInfra's official pricing page.