Medical speech benchmarkOctober 2026

Clinical accuracy.
Real-world speed.

The latest speech models, tested on the same medical audio. Compare batch and realtime transcription, including how quickly the text arrives.

1,513 English recordings · 7.2 hours · Speed for every API timed from the US (AWS Ohio), October 2026

Omi Medical v1.1 · Batch resultsHow we measure
0.87%Medical word errorLowest among 20 tested APIs
1 / 442Drug-term errors0.23% error rate
96.81%Dosage / unit matches91 of 94 reference events
131 msMedian response222 ms p95 · short clips

Batch transcription

The latest speech APIs alongside Omi’s open model, tested on the same medical audio.

Swipe the table to compare accuracy and speed.

Batch transcription, edition 2026-10-18.1. Medical WER ascending; ties ordered by dose / unit matches. Accuracy on 1,513 recordings.
#Model
1Omi Medical v1.10.87%25 / 2,8725.77%0.23%1 / 44296.81%91 / 943.19%3 / 9498.91%97.85%131 ms222 ms200
2Omi Medical-1 Flash1.04%30 / 2,8725.98%0.68%3 / 44296.81%91 / 943.19%3 / 9497.85%97.33%—Async only · see note—0
3Voxtral Small Transcribe 31.04%30 / 2,8725.58%0.68%3 / 44284.04%79 / 9415.96%15 / 9491.86%87.78%742 ms1,526 ms200
4ElevenLabs Scribe v2 Medical1.08%31 / 2,8725.87%0.45%2 / 44286.17%81 / 9413.83%13 / 9486.17%86.17%279 ms455 ms200
5Google Chirp 31.11%32 / 2,8726.31%1.13%5 / 44281.91%77 / 9418.09%17 / 9490.59%86.03%790 ms1,124 ms200
6ElevenLabs Scribe v21.15%33 / 2,8726.49%0.68%3 / 44281.91%77 / 9418.09%17 / 9486.52%84.15%417 ms683 ms200
7Microsoft MAI-Transcribe-21.18%34 / 2,8725.64%2.04%9 / 44285.11%80 / 9414.89%14 / 9489.89%87.43%200 ms474 ms200
8Speechmatics Enhanced Medical1.18%34 / 2,8726.37%0.90%4 / 44280.85%76 / 9419.15%18 / 9484.44%82.61%1,428 ms2,124 ms200
9Voxtral Mini Transcribe 31.22%35 / 2,8725.69%1.36%6 / 44279.79%75 / 9420.21%19 / 9490.36%84.75%603 ms1,120 ms200
10Deepgram Nova-3 Pharma1.22%35 / 2,8726.26%1.13%5 / 44280.85%76 / 9419.15%18 / 9493.83%86.86%373 ms876 ms200
11OpenAI GPT Transcribe1.32%38 / 2,8727.23%0.68%3 / 44279.79%75 / 9420.21%19 / 9488.24%83.80%584 ms992 ms200
12AssemblyAI Universal 3.5 Pro Medical1.50%43 / 2,8726.45%1.13%5 / 44285.11%80 / 9414.89%14 / 9484.21%84.66%2,692 ms3,876 ms200
13Soniox async v51.57%45 / 2,8726.63%2.26%10 / 44282.98%78 / 9417.02%16 / 9484.78%83.87%2,214 ms2,775 ms200
14Meta Muse Voice Transcribe 1.01.64%47 / 2,8726.43%1.13%5 / 44285.11%80 / 9414.89%14 / 9493.02%88.89%1,405 ms2,498 ms200
15Grok Voice Transcribe 2.01.74%50 / 2,8726.53%1.36%6 / 44276.60%72 / 9423.40%22 / 9492.31%83.72%323 ms663 ms200
16omi-medical-edge-1Open weightsEarlier edition · v7Download model ↗2.16%62 / 2,8726.64%4.52%20 / 44287.23%82 / 9412.77%12 / 9493.18%90.11%———
17Deepgram Nova-3 Medical2.26%65 / 2,8727.09%1.36%6 / 44273.40%69 / 9426.60%25 / 9488.46%80.23%347 ms674 ms200
18Gemini 3.5 Transcribe2.44%70 / 2,8727.89%1.36%6 / 44284.04%79 / 9415.96%15 / 9492.94%88.27%1,260 ms1,705 ms200
19Corti STT2.65%76 / 2,8728.06%2.94%13 / 44276.60%72 / 9423.40%22 / 9484.71%80.45%2,739 ms3,382 ms200
20Speechmatics Oak-12.92%84 / 2,8727.61%5.88%26 / 44275.53%71 / 9424.47%23 / 9489.87%82.08%1,371 ms1,675 ms200
21Amazon Transcribe Medical4.28%123 / 2,8726.19%2.26%10 / 44279.79%75 / 9420.21%19 / 9490.36%84.75%1,831 ms1,928 ms200

Speed includes upload, processing and response. 200 short clips, one pass. Speed for every API timed from the US (AWS Ohio), October 2026.

21 models

Open-model rows use earlier v7 accuracy results. Dosage scores use the current 94-event panel; comparable API speed is unavailable.

Omi Medical-1 Flash (omi-medical-1-flash) runs English async jobs in the US only, so it is not in the 200-clip speed test. On a single RTX PRO 6000, a 1-hour file is transcribed in a median 3.89 s (one request per file, median of 6 runs). About Flash

Lower is better for error rates and speed. Higher is better for dosage / unit matches. Medical-WER ties are ordered by dose / unit matches, highest first. Counts appear below each clinical metric. p95 is the delay within which 95% of measured responses arrive.

Additional result: AssemblyAI Sync Dictation

Short recordings only (1,465 clips), so this endpoint is separate from the full-dataset ranking. WER 5.31%; medical WER 1.19% (16 / 1,347); drug errors 0.77% (2 / 259); dosage / unit matches 81.82% (63 / 77). Median response 0.344 s; p95 0.567 s.

Same audio. Clear measurements.

Omi-run evaluation. Same recordings and scorer for every model; a fixed subset for speed. Accuracy and speed are measured separately.

What changed in Omi Medical v1.1

v1.1 is the October 2026 release of our medical transcription service. The API model ID stays omi-medical-1. The gains are in live transcription; batch medical-term and drug-name results are unchanged.

Realtime results on the same 1,513-recording English board
Measurev1 (September)v1.1
Medical-term errors38 / 2,87230 / 2,872
Drug-name errors5 / 4423 / 442
Overall WER6.55%6.17%
Dose / unit matches82 / 9490 / 94

Realtime speed is now measured from a shared end of speech, so we don't compare it with the earlier method.

Safeguards in v1.1

New guards help prevent stray text from digital silence and tiny audio fragments at the end of a turn. Improved speech detection helps recover speech masked by electrical hum; only speech detection is filtered, and the recognizer hears the original audio.

These reduce specific failure modes. Mumbled or heavily masked speech can still be misheard, so review medical transcripts before use.

Dataset and accuracy

1,513 English recordings, 7.2 hours. Every model uses the same audio and scorer. Medical WER covers 2,872 medical words; drug error covers 442 occurrences. Dosage / unit matching uses 94 reference events.

Omi accuracy comes from a US run on 17–18 October 2026 (api.us.omi.health; realtime recordings four streams at a time). Omi Medical-1 Flash accuracy comes from a US production run on 5 November 2026 (async jobs, two at a time). Other providers’ accuracy comes from September 2026; Speechmatics Oak-1 and MAI-Transcribe-2 Streaming were added in October 2026.

Realtime accuracy uses confirmed segments, including the final pieces at Stop. Drafts and Omi’s separate closing-record revisions are excluded. Scores describe the tested configurations, not clinical safety; small differences do not establish statistical significance.

How speed is measured

Where: batch APIs were timed on 18 October and realtime APIs on 19 October 2026 from one AWS client in Ohio (us-east-2), using each provider’s US endpoint where one is offered. Corti was reached through its EU endpoint, the region of our account. MAI-Transcribe-2 Streaming ran on an Azure Sweden Central deployment.

Sample: 200 short clips (<30 seconds), one pass, one request or stream at a time per model; models ran side by side. One warm-up request per model is excluded. Failed attempts are excluded, never replaced by retries.

Batch: request start to usable transcript, including upload and processing. Async endpoints are polled every 500 ms, so smaller differences between them are not meaningful.

Realtime: each clip is cut at a shared detected end of speech and streamed at normal speed. At that point the client sends the provider’s documented finish signal, and the clock runs until the last final text arrives. Accuracy is measured separately on all 1,513 recordings.

Scorer and citation

Scorer on GitHub · Diarization benchmark

Omi Health, Medical Speech-to-Text Benchmark. October 2026, edition 2026-10-18.1 (extends 2026-09-26.1).

Try it on your medical audio.

Start with the API. Compare the results that matter to your application.

Start building