whatmodel.sh / stt / GPT-Realtime-Whisper
GPT-Realtime-Whisper
↗ share
+ contribute
GPT-Realtime-Whisper is a closed-weights speech-to-text model from OpenAI, released May 2026. Licence: proprietary.
Catalogue identity notes
OpenAI streaming speech-to-text model, API alias/snapshot gpt-realtime-whisper (model page snapshot list shows only gpt-realtime-whisper; 'Sep 30, 2024 knowledge cutoff'). Released May 7, 2026 with GPT-Realtime-2 (OpenAI API changelog: 'GPT-Realtime-Whisper for streaming speech-to-text'). Closed weights; not the open Whisper checkpoints. Distinct from gpt-live-transcribe (2026-07-29), which Pipecat configured but has not published.
Score within the Open ASR English short-form WER (March 2026) comparison set only.
$ how it compares
The speech-to-text models ranked either side of GPT-Realtime-Whisper, out of 176 with a score, then more from OpenAI.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 46 |  Voxtral Mini Transcribe V2 | 59±3 | api $0.0030/audio min | ||
| 47 |  Avalon v1 (en) | 59±5 | comparison only | ||
| 48 | GPT-Realtime-Whisper | 59±3 | api $0.017/audio min | ||
| 49 |  Canary-1B | 59±3 | local | ||
| 50 |  Parakeet Ultra | 59±3 | local | ||
| more from OpenAI | |||||
| 13 | GPT Transcribe | 64±4 | api $0.0045/audio min | ||
| 61 | Whisper large-v3-turbo | 57±1 | api $0.0005/audio min +local | ||
| 67 | Whisper large-v3 | 57±1 | api $0.0018/audio min +local | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
other tests1 not in a task · 96% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Pipecat STT Benchmark · pooled semantic WER (1000 smart-turn samples, Sonnet 5.5 judge) | 2.3% ✎ | ▃·▃▅·▅▅·▃▅██ beats 5/16 | 96% |
$ where to run it
Through an API
api · $0.017/audio min · developers.openai.com ↗ · price sourceUSD per input audio minute: OpenAI pricing, Transcription models table, 'gpt-realtime-whisper | Live transcription | $0.017 / minute'; the model page states it 'is priced by audio duration rather than text tokens'. Product: Realtime API transcription. API model: gpt-realtime-whisper. Price basis: published per-minute rate. Currency: USD. OpenAI official pricing docs, observed 2026-10-09.
$ how the score was worked out · Pipecat STT Benchmark · seman… 96%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 Pipecat STT Benchmark · semantic WER results table (pipecat-ai/stt-benchmark README) | 1 | 96% |
| prior · pull toward the middle | 4% | |
| weighted average | = 59.0 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json