Live transcription just became a first-class api. openai's gpt-realtime-whisper, out this week,…
live transcription just became a first-class api. openai's gpt-realtime-whisper, out this week, streams captions and meeting notes as people speak instead of waiting for the recording to end.
the latency that made meeting bots feel broken is the product now.
Context
OpenAI's docs describe gpt-realtime-whisper as a streaming speech-to-text model for applications that need low-latency transcript deltas from live audio, designed for realtime use where developers tune latency and accuracy. It takes audio and text input and returns text, and its pricing page entry is $0.017 per minute of audio.
The realtime transcription guide says to use it when an application needs text from a microphone, call or other live audio stream without a spoken assistant response, and that the model returns transcript deltas as speech arrives and a final transcript when the application commits each audio turn.
The streaming behavior and the live audio use case match the docs. The pages read carry no release date, so 'out this week' is unsupported here, not refuted.
'Captions and meeting notes' are the author's examples. The docs describe microphone, call and live stream transcription, which covers them. 'First-class api' is the author's line, though a dedicated model ID and a guide do support it.
'The latency that made meeting bots feel broken' is the author's opinion. The pages read give no latency figures.
Related work
- GPT-Realtime-Whisper (OpenAI API docs) ↗Source for the model description, modalities and per minute price.
- Realtime transcription (OpenAI API docs) ↗Guide for realtime transcription with transcript deltas and final transcripts.
Watch next
- Find OpenAI's announcement for the release date. Look for independent latency and word error rate measurements.
Sources
- GPT-Realtime-Whisper (OpenAI API docs)developers.openai.com
- Realtime transcription (OpenAI API docs)developers.openai.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 9 October 2026 at 10:20 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →