PERIT-ALIGN · Leaderboard
Word alignment, measured against human experts
14 word-alignment models run on the same real-world audio and scored word by word against timestamps marked and reviewed by people.
| Rank | Model | Boundary MAE | Median boundary error | P90 boundary error | Start MAE | End MAE |
|---|---|---|---|---|---|---|
| 14 | nemotron-3.5-asr-streaming-multilingual-0.6bNVIDIA | 258.0 ms | 240.0 ms | 470.0 ms | 290.4 ms | 225.5 ms |
| 13 | parakeet-tdt-0.6b-v3NVIDIA | 116.8 ms | 80.0 ms | 240.0 ms | 85.3 ms | 148.4 ms |
| 12 | whisper-1OpenAI | 94.7 ms | 60.0 ms | 200.0 ms | 100.8 ms | 88.7 ms |
| 11 | whisper-large-v3-turboOpenAI | 97.8 ms | 60.0 ms | 220.0 ms | 110.3 ms | 85.4 ms |
By tolerance
Boundary error
Drift & shape
Coverage & speed
| ||||||
| 10 | mai-transcribe-2Microsoft | 66.6 ms | 50.0 ms | 140.0 ms | 84.9 ms | 48.3 ms |
| 09 | whisper-large-v3OpenAI | 90.1 ms | 49.6 ms | 169.4 ms | 98.6 ms | 81.6 ms |
| 08 | nova-3Deepgram | 76.6 ms | 80.0 ms | 160.0 ms | 68.8 ms | 84.5 ms |
| 07 | universal-3-5-proAssemblyAI | 68.6 ms | 52.0 ms | 136.0 ms | 70.8 ms | 66.4 ms |
| 06 | qwen3-asr-1.7bQwen | 70.0 ms | 0.0 ms | 180.0 ms | 102.1 ms | 37.8 ms |
| 05 | qwen3-asr-0.6bQwen | 68.2 ms | 0.0 ms | 160.0 ms | 98.6 ms | 37.8 ms |
| 04 | grok-stt-1.0xAI | 52.7 ms | 38.0 ms | 104.0 ms | 57.5 ms | 47.9 ms |
| 03 | voxtral-mini-transcribeMistral AI | 45.6 ms | 40.0 ms | 80.0 ms | 47.3 ms | 43.8 ms |
| 02 | transcribe-1Fish Audio | 24.5 ms | 0.0 ms | 80.0 ms | 25.0 ms | 23.9 ms |
| 01 | Perit Golden Data Fine-tuned ModelPerit | 23.8 ms | 16.0 ms | 47.0 ms | 22.9 ms | 24.6 ms |
Blue marks the best value in each column; a dash means the metric does not apply to that model. Open a model for all 19 metrics and where it places on each. Perit's model is a forced aligner: it places the words of a known transcript in time, while the other models transcribe the audio and time the words in one pass.Showing 14 of 14
Best on each axis
The winners, one question at a time.
- Most accurate timing
- 97.85%Timing Accuracy @100 msPerit Golden Data Fine-tuned ModelPerit
- Best speech-to-text timestamps
- 92.33%Timing Accuracy @100 mstranscribe-1Fish Audio
- Tightest timing
- 83.90%Timing Accuracy @50 msPerit Golden Data Fine-tuned ModelPerit
- Smallest boundary error
- 23.8 msBoundary MAEPerit Golden Data Fine-tuned ModelPerit
- Best span overlap
- 84.98%Mean IoUtranscribe-1Fish Audio
- Widest coverage
- 97.70%Word coveragemai-transcribe-2Microsoft
Glossary
What every column means.
Each model's words are matched to the reference words and compared boundary by boundary — start against start, end against end. Coverage and the transcript's own WER are reported next to the timing, so a model cannot look precise by timing only the easy words.
By tolerance
How many words and boundaries land inside each window, from 20 ms to 200 ms.
- Timing Accuracy @50 ms% · higher is better
- The same, at a 50 ms tolerance — how many words a model places tightly, not just roughly.
- Timing Accuracy @100 ms% · higher is better
- Headline metric. Share of reference words whose start and end both land within 100 ms of the reference timing.
- Boundaries within 20 ms% · higher is better
- Share of individual word boundaries (starts and ends counted separately) within 20 ms of the reference.
- Boundaries within 50 ms% · higher is better
- Share of word boundaries within 50 ms of the reference.
- Boundaries within 100 ms% · higher is better
- Share of word boundaries within 100 ms of the reference.
- Boundaries within 200 ms% · higher is better
- Share of word boundaries within 200 ms of the reference.
Boundary error
How far off the boundaries are, in milliseconds — the typical one, the worst tenth, starts against ends.
- Boundary MAEms · lower is better
- Mean absolute gap between predicted and reference word boundaries, starts and ends pooled. The tie-breaker in the ranking.
- Median boundary errorms · lower is better
- Boundary error of the typical boundary.
- P90 boundary errorms · lower is better
- Boundary error at the 90th percentile — how far off the worst tenth of boundaries land.
- Start MAEms · lower is better
- Mean absolute error of word start times alone.
- End MAEms · lower is better
- Mean absolute error of word end times alone.
Drift & shape
Whether a model runs systematically early or late, and whether its word spans have real length.
- Start biasms · closer to 0 is better
- Average signed offset of start times against the reference. Near zero means no systematic drift; the sign shows which way a model leans.
- End biasms · closer to 0 is better
- Average signed offset of end times against the reference.
- Mean IoU% · higher is better
- Average overlap between each word's predicted and reference time span (intersection over union).
- Zero-duration words% · lower is better
- Share of returned words with no length — start and end on the same timestamp.
Coverage & speed
How many words got a timestamp at all, how good the transcript underneath was, and request time — measured under concurrent load, so indicative only.
- Word coverage% · higher is better
- Share of reference words the model returned a timestamp for. Timing is only scored on words that are covered.
- Transcript WER% · lower is better
- Word error rate of the transcript the timestamps came with. Context only — it does not affect the ranking, and a forced aligner, which is given the transcript, has none.
- Success Rate% · higher is better
- Share of requests that returned timestamps.
- Median Latencys · lower is better
- Median request time. Measured under concurrent load; indicative only.
Want your model on this board?
We run it on the same audio, under the same scoring, and publish it next to the other 14.