Glean's Model Hub rate card adds real-time audio pricing and two new models
Glean'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.
GPT Realtime 1.5 and GPT Realtime 2 each listed as a single ambiguous row: $5.00 per million input tokens, $0.50 cache read, no output rate disclosed. No Deepgram or GPT-4o Mini TTS pricing on the card.
GPT Realtime 1.5, 2, and 2.1 each split into three modality-specific rows — Audio $32.00 in / $64.00 out, Text $4.00 in / $16.00-$24.00 out, Image $5.00 in — plus new rows for GPT-4o Mini TTS ($0.60 text in / $12.00 audio out) and Deepgram Nova-3 Multilingual ($0.0117/min output, 2 channels).
Glean’s Core Suite Model Hub Usage table — the one place Glean publishes actual dollar figures — moved again this week, four business days after its 2026-07-30 rate cuts on GPT 5.6 Terra and Luna. The table’s own “last updated” stamp moved from 7/30/2026 to 8/5/2026, and the page footer from Jul 10 to Aug 5, 2026.
This refresh didn’t touch existing rates; it restructured and expanded the card. GPT Realtime 1.5 and GPT Realtime 2, previously each a single row showing only a $5.00 input rate and no disclosed output price, are now split into three modality-specific lines apiece — Audio, Text, and Image — with GPT Realtime 2.1 (already listed in Glean’s Premium model-tier table since early August but not previously priced here) added as a third fully-priced Realtime model. The Audio rows are the first real-time audio token pricing Glean has ever disclosed: $32.00 per million input tokens and $64.00 per million output tokens, well above the $4.00/$16.00-24.00 Text rows on the same models. Glean also added a GPT-4o Mini TTS row ($0.60 text input, $12.00 per million output tokens for audio) and its first Deepgram entry — Nova-3 Multilingual (2 channels) at $0.0117 per minute of output, the only per-minute rather than per-token row on the card.
Glean still discloses no seat price, no FlexCredit dollar value, and no Flexible Model Management percentage on either Enterprise Flex or Glean Core Suite; glean.com/pricing continues to 301-redirect to the homepage. The Enterprise Flex FlexCredit rate card and model-tier table are unchanged since the 2026-08-04 check.
Glean's usage-limits docs add a fourth cap tier: admins can now set a monthly usage limit per department (not pooled — the same cap applies to every member), enforced between an individual per-user override and the org-wide default user limit. Requires department metadata from the identity provider or org chart; Glean's own guidance recommends alert-only limits before enabling hard caps.