Pipecat Speech-to-Text Benchmark
Speech-to-text services transcribe the same 1,000 realtime audio clips. Semantic word error rate measures accuracy, and time to final segment measures streaming latency.
Pipecat publishes 24 model rows from 15 vendors, with pooled semantic WER and TTFS median, P95 and P99. The BSD-2-Clause repository has 143 GitHub stars and the dataset has 15 Hugging Face likes.
- Made by
- Pipecat
- Released
- August 2026
- Kind
- Speech recognition and translation
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech
- Last model added
- September 2026
- Scores refreshed
- November 21st, 2026
Muse Voice Transcribe 1.0 leads with 0.83% among models released by September 2026 Muse Voice Transcribe 1.0 leads with 0.83% among models released by September 2026
Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Pooled semantic word error rate from 1,000 realtime clips 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. 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
Muse Voice Transcribe 1.0 0.83%
- 2 AssemblyAI universal-3-6-pro 0.96%
- 3 Speechmatics linden-1 1.05%
- 4 Speechmatics 1.07%
- 5 Azure 1.18%
- 6 AssemblyAI universal-3-5-pro 1.22%
- 7 Cartesia ink-2 1.25%
- 8 Soniox stt-rt-v5 1.27%
- 9 Soniox stt-rt-v4 1.29%
- 10 AssemblyAI u3-rt-pro 1.34%
- 11 Deepgram nova-3-general 1.62%
- 12 AWS 1.75%
- 13 NVIDIA Nemotron 3.0 ASR (en) 1.95%
- 14
Gemini 3.5 Transcribe Live 2.24%
- 15 Smallest AI pulse 2.37%
- 16 GPT-Realtime-Whisper 2.73%
- 17 Google latest-long 2.85%
- 18 AssemblyAI universal-streaming-english 3.02%
- 19 gpt-4o-transcribe 3.06%
- 20 ElevenLabs scribe_v2_realtime 3.12%
- 21 Gradium default 3.71%
- 22 Cartesia ink-whisper 4.36%
- 23 NVIDIA Nemotron 3.5 ASR (multilingual) 4.58%
- 24 Mistral voxtral-mini-transcribe-realtime-2602 4.97%
Source: Official Pipecat Speech-to-Text Benchmark leaderboard
- 1
Muse Voice Transcribe 1.0 0.83%
- 2
Gemini 3.5 Transcribe Live 2.24%
- 3 GPT-Realtime-Whisper 2.73%
- 4 gpt-4o-transcribe 3.06%
Source: Official Pipecat Speech-to-Text Benchmark leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.83% | Sep 2026 | Official leaderboard | |
| 2 | AssemblyAI universal-3-6-pro | 0.96% | Official leaderboard | ||
| 3 | Speechmatics linden-1 | 1.05% | Official leaderboard | ||
| 4 | Speechmatics | 1.07% | Official leaderboard | ||
| 5 | Azure | 1.18% | Official leaderboard | ||
| 6 | AssemblyAI universal-3-5-pro | 1.22% | Official leaderboard | ||
| 7 | Cartesia ink-2 | 1.25% | Official leaderboard | ||
| 8 | Soniox stt-rt-v5 | 1.27% | Official leaderboard | ||
| 9 | Soniox stt-rt-v4 | 1.29% | Official leaderboard | ||
| 10 | AssemblyAI u3-rt-pro | 1.34% | Official leaderboard | ||
| 11 | Deepgram nova-3-general | 1.62% | Official leaderboard | ||
| 12 | AWS | 1.75% | Official leaderboard | ||
| 13 | NVIDIA Nemotron 3.0 ASR (en) | 1.95% | Official leaderboard | ||
| 14 | | 2.24% | Aug 2026 | Official leaderboard | |
| 15 | Smallest AI pulse | 2.37% | Official leaderboard | ||
| 16 | GPT-Realtime-Whisper | 2.73% | May 2026 | Official leaderboard | |
| 17 | Google latest-long | 2.85% | Official leaderboard | ||
| 18 | AssemblyAI universal-streaming-english | 3.02% | Official leaderboard | ||
| 19 | gpt-4o-transcribe | 3.06% | Mar 2025 | Official leaderboard | |
| 20 | ElevenLabs scribe_v2_realtime | 3.12% | Official leaderboard | ||
| 21 | Gradium default | 3.71% | Official leaderboard | ||
| 22 | Cartesia ink-whisper | 4.36% | Official leaderboard | ||
| 23 | NVIDIA Nemotron 3.5 ASR (multilingual) | 4.58% | Official leaderboard | ||
| 24 | Mistral voxtral-mini-transcribe-realtime-2602 | 4.97% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.83% | Sep 2026 | Official leaderboard | |
| 2 | | 2.24% | Aug 2026 | Official leaderboard | |
| 3 | GPT-Realtime-Whisper | 2.73% | May 2026 | Official leaderboard | |
| 4 | gpt-4o-transcribe | 3.06% | Mar 2025 | Official leaderboard |
3 models have held number one on the official leaderboard since its release in August 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Pipecat 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 Pipecat 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
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (24 entries).
- Pass Models scored 4 models we track, plus 20 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 pipecat.ai: 4 rows for models we track and 20 other entries at the last read.
- Pass Last read November 21st, 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.