OpenAIMistral AIAudio matchupSpeech to text
GPT-Realtime-Whispervs.Voxtral Mini Transcribe Realtime
Which model is better?
GPT-Realtime-Whisper is slightly ahead in independent tests, but the lead is not certain.
Pick GPT-Realtime-Whisper for
Pick Voxtral Mini Transcribe Realtime 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”.
Coval
GPT-Realtime-Whisper slightly aheadGPT-Realtime-Whisper is clearly ahead in 0 of 16 subtests and slightly in 7, Voxtral Mini Transcribe Realtime in 0 clearly and slightly in 2. 7 are level.
Coval Time to First Token (TTFT) maximum
lower is better, as of November 5, 2026
Coval RTF maximum
lower is better, as of November 5, 2026
Coval Word Error Rate (WER) per-sample maximum
lower is better, as of November 5, 2026
Show all 16 subtests
| Subtest | GPT-Realtime-Whisper | Voxtral Mini Transcribe Realtime | Verdict |
|---|---|---|---|
| Coval Time to First Token (TTFT) maximum (lower is better) | 10,968 ms | 40,858 ms | GPT-Realtime-Whisper slightly ahead |
| Coval RTF maximum (lower is better) | 5.55 | 15.59 | GPT-Realtime-Whisper slightly ahead |
| Coval Word Error Rate (WER) per-sample maximum (lower is better) | 100.0 % | 191.7 % | GPT-Realtime-Whisper slightly ahead |
| Coval Time to Final Segment (TTFS) maximum (lower is better) | 20,386 ms | 35,330 ms | GPT-Realtime-Whisper slightly ahead |
| Coval Time to Final Segment (TTFS) p50 (lower is better) | 547 ms | 339 ms | Voxtral Mini Transcribe Realtime slightly ahead |
| Coval AudioToFinal maximum (lower is better) | 30.01 s | 43.02 s | GPT-Realtime-Whisper slightly ahead |
| Coval pooled corpus Word Error Rate (WER) (lower is better) | 4.0 % | 5.4 % | GPT-Realtime-Whisper slightly ahead |
| Coval Time to Final Segment (TTFS) mean (lower is better) | 564 ms | 444 ms | Voxtral Mini Transcribe Realtime slightly ahead |
| Coval Word Error Rate (WER) per-sample mean (lower is better) | 4.5 % | 5.7 % | GPT-Realtime-Whisper slightly ahead |
| Coval Time to First Token (TTFT) mean (lower is better) | 1,820 ms | 1,872 ms | level |
| Coval AudioToFinal p50 (lower is better) | 8.88 s | 8.70 s | level |
| Coval RTF p50 (lower is better) | 1.07 | 1.05 | level |
| Coval Time to First Token (TTFT) p50 (lower is better) | 1,878 ms | 1,846 ms | level |
| Coval RTF mean (lower is better) | 1.09 | 1.07 | level |
| Coval AudioToFinal mean (lower is better) | 8.51 s | 8.38 s | level |
| Coval Word Error Rate (WER) per-sample p50 (lower is better) | 0.0 % | 0.0 % | level |
Source · latest November 5, 2026
Pipecat STT Benchmark
GPT-Realtime-Whisper slightly aheadGPT-Realtime-Whisper is clearly ahead in 0 of 2 subtests and slightly in 2, Voxtral Mini Transcribe Realtime in none.
Pipecat STT Benchmark semantic pooled WER
lower is better, as of October 16, 2026
Pipecat STT Benchmark semantic WER mean
lower is better, as of October 16, 2026
Source · latest October 16, 2026
Artificial Analysis
GPT-Realtime-Whisper slightly aheadArtificial Analysis AA-WER Streaming (final transcription)
lower is better, as of October 15, 2026
Source · latest October 15, 2026
All data side by side
Prices are verbatim from the provider, because audio models bill per minute, character or token.
| GPT-Realtime-Whisper | Voxtral Mini Transcribe Realtime | |
|---|---|---|
| Task | Speech to text | Speech to text |
| Developer | OpenAI | Mistral AI |
| Status | Active | Active |
| Weights | proprietary | open |
| Runs locally | no | yes |
| Price | $0.017 per audio minute | n/a |
| Popularity (out of 100) | 34 | 18 |