OpenAIElevenLabsAudio matchupSpeech to text
GPT-Transcribevs.Scribe v2
Which model is better?
Scribe v2 is slightly ahead in independent tests, but the lead is not certain.
Pick GPT-Transcribe for
Pick Scribe v2 for
Shared tests
Independent evaluators and leaderboards only, the newest shared snapshot per evaluator. With 95 % ranges a lead counts when they do not overlap; for hit rates, when it holds in a 95 % test. Without any uncertainty figure at most “slightly ahead”.
Hugging Face Open ASR
Scribe v2 slightly aheadScribe v2 is clearly ahead in 0 of 12 subtests and slightly in 9, GPT-Transcribe in 0 clearly and slightly in 2. 1 is level.
Open ASR English short-form WER: Voxpopuli-AA-Cleaned
lower is better, as of October 28, 2026
Open ASR English short-form WER: LS Other
lower is better, as of October 28, 2026
Open ASR English short-form WER: URGENT24
lower is better, as of October 28, 2026
Show all 12 subtests
| Subtest | GPT-Transcribe | Scribe v2 | Verdict |
|---|---|---|---|
| Open ASR English short-form WER: Voxpopuli-AA-Cleaned (lower is better) | 2.7 % | 1.4 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: LS Other (lower is better) | 3.4 % | 2.3 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: URGENT24 (lower is better) | 12.9 % | 9.2 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: Earnings22-Cleaned-AA-chunked (lower is better) | 6.6 % | 4.8 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: SPGISpeech (lower is better) | 2.2 % | 3.0 % | GPT-Transcribe slightly ahead |
| Open ASR English short-form WER: LS Clean (lower is better) | 1.3 % | 1.1 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: Average WER (11-set macro-average) (lower is better) | 6.0 % | 5.2 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: Voice Arena Monsoon (lower is better) | 3.6 % | 3.3 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: AMI-Cleaned (lower is better) | 9.0 % | 8.2 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: Private (scripted) (lower is better) | 3.2 % | 3.0 % | Scribe v2 slightly ahead |
| Open ASR English short-form WER: Private (conversational) (lower is better) | 12.9 % | 13.7 % | GPT-Transcribe slightly ahead |
| Open ASR English short-form WER: Gigaspeech-Cleaned (lower is better) | 7.8 % | 7.6 % | level |
Source · latest October 28, 2026
Appen MedTerm-90
levelAppen MedTerm-90 overall clinical entity recall
95 % ranges 80.7 % to 82.8 % and 82.6 % to 84.1 %, 30,017 tasks, as of October 16, 2026
Source · latest October 16, 2026
Artificial Analysis
Scribe v2 slightly aheadArtificial Analysis AA-WER (non-streaming)
lower is better, as of October 15, 2026
Source · latest October 15, 2026
Vendor figures
Figures from vendor reports that cover both models. They do not count toward the verdict.
Show 2 vendor figures
| Figure | Source | GPT-Transcribe | Scribe v2 |
|---|---|---|---|
| ElevenLabs MedTerm medical-term recall | ElevenLabs model evaluation | 74.0 % | 77.1 % |
| ElevenLabs MedTerm medical-term WER (lower is better) | ElevenLabs model evaluation | 12.3 % | 10.7 % |
All data side by side
Prices are verbatim from the provider, because audio models bill per minute, character or token.
| GPT-Transcribe | Scribe v2 | |
|---|---|---|
| Task | Speech to text | Speech to text |
| Developer | OpenAI | ElevenLabs |
| Status | Active | Active |
| Released | July 28, 2026 | n/a |
| Weights | proprietary | proprietary |
| Runs locally | no | no |
| Price | $0.0045 per audio minute | n/a |
| Popularity (out of 100) | 32 | 39 |