Medical Speech-to-Text Benchmark
Speech models transcribe 1,513 English medical recordings. A common scorer measures word errors, medical-term errors, drug-name errors, and dosage matches. Separate tests measure batch response time and final streaming latency.
The October edition extends the September 2026 benchmark. Accuracy uses 7.2 hours of audio, while timing uses 200 short clips. Failed timing attempts are excluded. The maker also sells a model on the board, and these scores do not establish clinical safety.
- Made by
- Omi Health
- Released
- September 2026
- Kind
- Speech recognition and translation
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Open-ended
- Measures
- Capability
- Subject
- Clinical medicine
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- New versions replace it
- Run by
- A host
- Based on
- A version of another benchmark
- In views
- Audio and speech, Biology and medicine
- Last model added
- September 2026
- Scores refreshed
- November 18th, 2026
Omi Medical v1.1 tops the leaderboard with 0.87% 2 models share the top score of 1.32% among models released by September 2026
Best score per entry from the benchmark's own leaderboard, read on November 18th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is GPT-Live-Transcribe, at 1.32%. This view shows only models we track, each at its best.
The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 Omi Medical v1.1 batch 0.87%
- 2 Omi Medical-1 Flash batch 1.04%
- 3 Voxtral Small Transcribe 3 batch 1.04%
- 4 ElevenLabs Scribe v2 Medical batch 1.08%
- 5 Google Chirp 3 batch 1.11%
- 6 ElevenLabs Scribe v2 batch 1.15%
- 7 Microsoft MAI-Transcribe-2 batch 1.18%
- 8 Speechmatics Enhanced Medical batch 1.18%
- 9 Deepgram Nova-3 Pharma batch 1.22%
- 10 Voxtral Mini Transcribe 3 batch 1.22%
- 11 Speechmatics Linden-1 realtime 1.29%
- 12 Microsoft MAI-Transcribe-2 Streaming realtime 1.32%
- 13 GPT-Live-Transcribe realtime 1.32%
- 14 GPT-Transcribe batch 1.32%
- 15 Voxtral Mini Transcribe Realtime 3 realtime 1.36%
- 16 AssemblyAI Universal 3.6 Pro realtime 1.39%
- 17 ElevenLabs Scribe v2 Realtime realtime 1.39%
- 18 Soniox v5 realtime 1.43%
- 19 AssemblyAI Universal 3.5 Pro Medical batch 1.50%
- 20 Soniox async v5 batch 1.57%
- 21
Muse Voice Transcribe 1.0 realtime 1.60%
- 22 Grok Voice Transcribe 2.0 batch 1.74%
- 23 omi-medical-edge-1 Open weights Earlier edition · v7 Download model ↗ batch 2.16%
- 24 Deepgram Nova-3 Medical batch 2.26%
- 25
Gemini 3.5 Transcribe batch 2.44%
- 26 Corti STT batch 2.65%
- 27 Speechmatics Oak-1 batch 2.92%
- 28
Gemini 3.5 Transcribe Live realtime 3.34%
- 29 Amazon Transcribe Medical batch 4.28%
Source: Official Medical Speech-to-Text Benchmark leaderboard
- 1 GPT-Live-Transcribe realtime 1.32%
- 2 GPT-Transcribe batch 1.32%
- 3
Muse Voice Transcribe 1.0 realtime 1.60%
- 4
Gemini 3.5 Transcribe batch 2.44%
- 5
Gemini 3.5 Transcribe Live realtime 3.34%
Source: Official Medical Speech-to-Text Benchmark leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Omi Medical v1.1 | 0.87% | batch | Official leaderboard | |
| 2 | Omi Medical-1 Flash | 1.04% | batch | Official leaderboard | |
| 3 | Voxtral Small Transcribe 3 | 1.04% | batch | Official leaderboard | |
| 4 | ElevenLabs Scribe v2 Medical | 1.08% | batch | Official leaderboard | |
| 5 | Google Chirp 3 | 1.11% | batch | Official leaderboard | |
| 6 | ElevenLabs Scribe v2 | 1.15% | batch | Official leaderboard | |
| 7 | Microsoft MAI-Transcribe-2 | 1.18% | batch | Official leaderboard | |
| 8 | Speechmatics Enhanced Medical | 1.18% | batch | Official leaderboard | |
| 9 | Deepgram Nova-3 Pharma | 1.22% | batch | Official leaderboard | |
| 10 | Voxtral Mini Transcribe 3 | 1.22% | batch | Official leaderboard | |
| 11 | Speechmatics Linden-1 | 1.29% | realtime | Official leaderboard | |
| 12 | Microsoft MAI-Transcribe-2 Streaming | 1.32% | realtime | Official leaderboard | |
| 13 | GPT-Live-Transcribe | 1.32% | realtime | Jul 2026 | Official leaderboard |
| 14 | GPT-Transcribe | 1.32% | batch | Jul 2026 | Official leaderboard |
| 15 | Voxtral Mini Transcribe Realtime 3 | 1.36% | realtime | Official leaderboard | |
| 16 | AssemblyAI Universal 3.6 Pro | 1.39% | realtime | Official leaderboard | |
| 17 | ElevenLabs Scribe v2 Realtime | 1.39% | realtime | Official leaderboard | |
| 18 | Soniox v5 | 1.43% | realtime | Official leaderboard | |
| 19 | AssemblyAI Universal 3.5 Pro Medical | 1.50% | batch | Official leaderboard | |
| 20 | Soniox async v5 | 1.57% | batch | Official leaderboard | |
| 21 | | 1.60% | realtime | Sep 2026 | Official leaderboard |
| 22 | Grok Voice Transcribe 2.0 | 1.74% | batch | Official leaderboard | |
| 23 | omi-medical-edge-1 Open weights Earlier edition · v7 Download model ↗ | 2.16% | batch | Official leaderboard | |
| 24 | Deepgram Nova-3 Medical | 2.26% | batch | Official leaderboard | |
| 25 | | 2.44% | batch | Aug 2026 | Official leaderboard |
| 26 | Corti STT | 2.65% | batch | Official leaderboard | |
| 27 | Speechmatics Oak-1 | 2.92% | batch | Official leaderboard | |
| 28 | | 3.34% | realtime | Aug 2026 | Official leaderboard |
| 29 | Amazon Transcribe Medical | 4.28% | batch | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-Live-Transcribe | 1.32% | realtime | Jul 2026 | Official leaderboard |
| 2 | GPT-Transcribe | 1.32% | batch | Jul 2026 | Official leaderboard |
| 3 | | 1.60% | realtime | Sep 2026 | Official leaderboard |
| 4 | | 2.44% | batch | Aug 2026 | Official leaderboard |
| 5 | | 3.34% | realtime | Aug 2026 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Medical Speech-to-Text Benchmark leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Medical Speech-to-Text Benchmark leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Audio and video
- Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy.
Qwen 3.8 Max 88.7%
- LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization.
Gemini 4 Argon 91.7%
- MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy.
Qwen 3.8 Max 82.4%
- MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy.
Qwen 3.8 Max 90.8%
- FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better.
Phi-4-multimodal 4.00%
- VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules.
Qwen3.5-Omni-Plus 93.1%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (29 entries).
- Pass Models scored 5 models we track, plus 24 other entries on its leaderboard.
- Pass Newest score The newest is from September 2026, for Muse Voice Transcribe 1.0.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
- Pass Importer We read the board on omi.health: 6 rows for models we track and 24 other entries at the last read.
- Pass Last read November 18th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.