ElevenLabsDeepgramAudio matchupSpeech to text
Scribe v2vs.Nova-3
Which model is better?
Scribe v2 is slightly ahead in independent tests, but the lead is not certain.
Pick Scribe v2 for
Pick Nova-3 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”.
Nyra Verbatim Speech Benchmark
Scribe v2 slightly aheadScribe v2 is clearly ahead in 0 of 12 subtests and slightly in 12, Nova-3 in none.
Nyra Verbatim Speech Benchmark (English) Vocal Sound F1
4,957 tasks, as of October 18, 2026
Nyra Verbatim Speech Benchmark (German) Vocal Sound F1
202 tasks, as of October 18, 2026
Nyra Verbatim Speech Benchmark (German) Filler F1
202 tasks, as of October 18, 2026
Show all 12 subtests
| Subtest | Scribe v2 | Nova-3 | Verdict |
|---|---|---|---|
| Nyra Verbatim Speech Benchmark (English) Vocal Sound F1 | 83.4 % | 0.0 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) Vocal Sound F1 | 80.4 % | 0.0 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) Filler F1 | 98.5 % | 1.0 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) Cutoff F1 | 77.7 % | 14.1 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) vWER (lower is better) | 4.2 % | 19.2 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) Disfluency F1 | 89.3 % | 20.4 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (English) Filler F1 | 95.5 % | 45.7 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (English) vWER (lower is better) | 3.2 % | 6.3 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (English) Disfluency F1 | 90.3 % | 57.3 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (English) Cutoff F1 | 80.1 % | 72.2 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (German) Rep F1 | 88.1 % | 79.8 % | Scribe v2 slightly ahead |
| Nyra Verbatim Speech Benchmark (English) Rep F1 | 87.9 % | 83.2 % | Scribe v2 slightly ahead |
Source · latest October 18, 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
All data side by side
Prices are verbatim from the provider, because audio models bill per minute, character or token.