Benchmark Almanac

SPGISpeech

zoom/scribe_v2.1_pro 1.39% tops its leaderboard Voxtral Small 1.84% leads the models we track
Not in model cards

Models transcribe English financial recordings. The references include formatting such as capitalization and punctuation.

Made by
Kensho and collaborators
Released
April 2021
Kind
Speech recognition and translation
Level
Everyday
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Audio → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
Last model added
October 2026
Scores refreshed
November 20th, 2026
68
Entries on Open ASR Leaderboard’s leaderboard, 9 models we track

zoom/scribe_v2.1_pro tops the leaderboard with 1.39% Voxtral Small leads with 1.84%

Best score per entry from Open ASR Leaderboard’s leaderboard, read on November 20th, 2026. 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. The best model we track is Voxtral Small, at 1.84%. This view shows only models we track, each at its best.

  1. 1 zoom/scribe_v2.1_pro 1.39%
  2. 2 OtterAI/speech-v3.1 1.54%
  3. 3 TheStageAI/thewhisper-large-v3-turbo 1.62%
  4. 4 soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, transducer_modified_beam_search) 1.64%
  5. 5 opusclip/steno-v1-preview 1.66%
  6. 6 nvidia/canary-qwen-2.5b 1.70%
  7. 7 OpenMOSS-Team/MOSS-Transcribe-Diarize 1.70%
  8. 8 HojoAI/Hojo-ASR-V1 1.74%
  9. 9 ibm-granite/granite-speech-5.0-470m-turboctc-nc 1.75%
  10. 10 assemblyai/universal-3-5-pro 1.81%
  11. 11 Voxtral Small 1.84%
  12. 12 Musci-research/Musci-ASR-qwen-1019-hf 1.86%
  13. 13 nvidia/parakeet-tdt-0.6b-v2 (fast-gpu-asr) 1.87%
  14. 14 microsoft/azure-speech-09-2026 1.90%
  15. 15 nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8 1.90%
  16. 16 smallestai/pulse 1.91%
  17. 17 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 1.91%
  18. 18 nvidia/parakeet-tdt-0.6b-v2 1.94%
  19. 19 nvidia/canary-1b-v2 2.03%
  20. 20 FunAudioLLM/Fun-ASR-Nano-2512-hf 2.05%
  21. 21 modulate/multilingual 2.06%
  22. 22 bosonai/higgs-audio-v3-stt 2.07%
  23. 23 sophea/asr-k1 2.11%
  24. 24 mistralai/Voxtral-Mini-3B-2507 2.15%
  25. 25 sprag/symphony 2.21%
  26. 26 p99lab/audrey-0.7 2.22%
  27. 27 Voxtral Mini Transcribe Realtime (2602) 2.23%
  28. 28 GPT-Transcribe 2.23%
  29. 29 espnet/owsm_ctc_v4_1B 2.31%
  30. 30 FluidVoice-Labs/fluid-speech-pro 2.31%
  31. 31 salesforce-sovereign-intelligence/SOV-Transcribe-1 2.40%
  32. 32 Edge0/ARK-ASR-3B 2.46%
  33. 33 fonada/kapture-v3 2.46%
  34. 34 OpenLLM-France/Luciole-8B-Audio-1.0 2.53%
  35. 35 Qwen/Qwen3-ASR-1.7B-hf 2.57%
  36. 36 abr-ai/niagara-84m-batch.en 2.61%
  37. 37 yoob/yoob-asr-pro-v1 2.61%
  38. 38 Edge0/ARK-ASR-0.6B 2.63%
  39. 39 gladia/solaria-3 2.68%
  40. 40 nvidia/nemotron-speech-streaming-en-0.6b 2.69%
  41. 41 Whisper large-v3 2.71%
  42. 42 reson8/rumble-2 2.73%
  43. 43 oruk/orukqwen-asr-1.7b 2.74%
  44. 44 Qwen/Qwen3-ASR-0.6B-hf 2.74%
  45. 45 Cohere Transcribe 2.74%
  46. 46 soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, ctc_greedy_search) 2.75%
  47. 47 Whisper large-v3-turbo 2.79%
  48. 48 Banafo/Kroko-ASR-XL-EN-Offline (fast-gpu-asr) 2.99%
  49. 49 elevenlabs/scribe_v2 3.02%
  50. 50 Reson8/Rumble-2-Flash 3.08%
  51. 51 Muse Voice Transcribe 1.0 3.21%
  52. 52 hf-audio/omniasr-llm-7b-v2 3.22%
  53. 53 ibm-granite/granite-speech-5.0-470m-turboctc 3.22%
  54. 54 ibm-granite/granite-speech-4.1-2b-nar 3.23%
  55. 55 nvidia/nemotron-3.5-asr-streaming-0.6b 3.27%
  56. 56 FermionResearch/Phonon-2.5 3.30%
  57. 57 FermionResearch/Phonon-2.5-English 3.32%
  58. 58 ibm-granite/granite-speech-5.0-2b 3.36%
  59. 59 ibm-granite/granite-speech-4.1-2b 3.47%
  60. 60 microsoft/VibeVoice-ASR-HF 3.47%
  61. 61 ibm-granite/granite-4.0-1b-speech 3.55%
  62. 62 nvidia/parakeet-tdt-0.6b-v3 (fast-gpu-asr) 3.56%
  63. 63 nvidia/parakeet-tdt-0.6b-v3 3.63%
  64. 64 Edge0/Audio8-ASR-0.1B 3.72%
  65. 65 futo-org/asr4all-l 4.71%
  66. 66 hf-audio/omniasr-ctc-7b-v2 5.16%
  67. 67 Gemma 4 E4B 5.49%
  68. 68 qvac/TranscribePsy-250M-EU-sub 8.39%

Source: Open ASR Leaderboard’s leaderboard

#EntryScoreSettingDateLinkSource
1 zoom/scribe_v2.1_pro 1.39% zoom.ai Official leaderboard
2 OtterAI/speech-v3.1 1.54% otter.ai Official leaderboard
3 TheStageAI/thewhisper-large-v3-turbo 1.62% huggingface.co Official leaderboard
4 soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, transducer_modified_beam_search) 1.64% github.com Official leaderboard
5 opusclip/steno-v1-preview 1.66% platform.labs.opus.pro Official leaderboard
6 nvidia/canary-qwen-2.5b 1.70% huggingface.co Official leaderboard
7 OpenMOSS-Team/MOSS-Transcribe-Diarize 1.70% huggingface.co Official leaderboard
8 HojoAI/Hojo-ASR-V1 1.74% huggingface.co Official leaderboard
9 ibm-granite/granite-speech-5.0-470m-turboctc-nc 1.75% huggingface.co Official leaderboard
10 assemblyai/universal-3-5-pro 1.81% assemblyai.com Official leaderboard
11 Voxtral Small 1.84% Jul 2025 huggingface.co Official leaderboard
12 Musci-research/Musci-ASR-qwen-1019-hf 1.86% huggingface.co Official leaderboard
13 nvidia/parakeet-tdt-0.6b-v2 (fast-gpu-asr) 1.87% github.com Official leaderboard
14 microsoft/azure-speech-09-2026 1.90% learn.microsoft.com Official leaderboard
15 nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8 1.90% huggingface.co Official leaderboard
16 smallestai/pulse 1.91% smallest.ai Official leaderboard
17 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 1.91% Oct 2026 huggingface.co Official leaderboard
18 nvidia/parakeet-tdt-0.6b-v2 1.94% huggingface.co Official leaderboard
19 nvidia/canary-1b-v2 2.03% huggingface.co Official leaderboard
20 FunAudioLLM/Fun-ASR-Nano-2512-hf 2.05% huggingface.co Official leaderboard
21 modulate/multilingual 2.06% docs.modulate.ai Official leaderboard
22 bosonai/higgs-audio-v3-stt 2.07% huggingface.co Official leaderboard
23 sophea/asr-k1 2.11% platform.sophea.ai Official leaderboard
24 mistralai/Voxtral-Mini-3B-2507 2.15% huggingface.co Official leaderboard
25 sprag/symphony 2.21% sprag.ai Official leaderboard
26 p99lab/audrey-0.7 2.22% p99lab.com Official leaderboard
27 Voxtral Mini Transcribe Realtime (2602) 2.23% Feb 2026 huggingface.co Official leaderboard
28 GPT-Transcribe 2.23% Jul 2026 platform.openai.com Official leaderboard
29 espnet/owsm_ctc_v4_1B 2.31% huggingface.co Official leaderboard
30 FluidVoice-Labs/fluid-speech-pro 2.31% huggingface.co Official leaderboard
31 salesforce-sovereign-intelligence/SOV-Transcribe-1 2.40% huggingface.co Official leaderboard
32 Edge0/ARK-ASR-3B 2.46% huggingface.co Official leaderboard
33 fonada/kapture-v3 2.46% fonadalabs.ai Official leaderboard
34 OpenLLM-France/Luciole-8B-Audio-1.0 2.53% huggingface.co Official leaderboard
35 Qwen/Qwen3-ASR-1.7B-hf 2.57% huggingface.co Official leaderboard
36 abr-ai/niagara-84m-batch.en 2.61% huggingface.co Official leaderboard
37 yoob/yoob-asr-pro-v1 2.61% yoob.com Official leaderboard
38 Edge0/ARK-ASR-0.6B 2.63% huggingface.co Official leaderboard
39 gladia/solaria-3 2.68% gladia.io Official leaderboard
40 nvidia/nemotron-speech-streaming-en-0.6b 2.69% huggingface.co Official leaderboard
41 Whisper large-v3 2.71% Nov 2023 huggingface.co Official leaderboard
42 reson8/rumble-2 2.73% reson8.dev Official leaderboard
43 oruk/orukqwen-asr-1.7b 2.74% huggingface.co Official leaderboard
44 Qwen/Qwen3-ASR-0.6B-hf 2.74% huggingface.co Official leaderboard
45 Cohere Transcribe 2.74% Mar 2026 huggingface.co Official leaderboard
46 soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, ctc_greedy_search) 2.75% github.com Official leaderboard
47 Whisper large-v3-turbo 2.79% Oct 2024 huggingface.co Official leaderboard
48 Banafo/Kroko-ASR-XL-EN-Offline (fast-gpu-asr) 2.99% huggingface.co Official leaderboard
49 elevenlabs/scribe_v2 3.02% elevenlabs.io Official leaderboard
50 Reson8/Rumble-2-Flash 3.08% huggingface.co Official leaderboard
51 Muse Voice Transcribe 1.0 3.21% Sep 2026 developer.meta.com Official leaderboard
52 hf-audio/omniasr-llm-7b-v2 3.22% huggingface.co Official leaderboard
53 ibm-granite/granite-speech-5.0-470m-turboctc 3.22% huggingface.co Official leaderboard
54 ibm-granite/granite-speech-4.1-2b-nar 3.23% huggingface.co Official leaderboard
55 nvidia/nemotron-3.5-asr-streaming-0.6b 3.27% huggingface.co Official leaderboard
56 FermionResearch/Phonon-2.5 3.30% huggingface.co Official leaderboard
57 FermionResearch/Phonon-2.5-English 3.32% huggingface.co Official leaderboard
58 ibm-granite/granite-speech-5.0-2b 3.36% huggingface.co Official leaderboard
59 ibm-granite/granite-speech-4.1-2b 3.47% huggingface.co Official leaderboard
60 microsoft/VibeVoice-ASR-HF 3.47% huggingface.co Official leaderboard
61 ibm-granite/granite-4.0-1b-speech 3.55% huggingface.co Official leaderboard
62 nvidia/parakeet-tdt-0.6b-v3 (fast-gpu-asr) 3.56% github.com Official leaderboard
63 nvidia/parakeet-tdt-0.6b-v3 3.63% huggingface.co Official leaderboard
64 Edge0/Audio8-ASR-0.1B 3.72% huggingface.co Official leaderboard
65 futo-org/asr4all-l 4.71% huggingface.co Official leaderboard
66 hf-audio/omniasr-ctc-7b-v2 5.16% huggingface.co Official leaderboard
67 Gemma 4 E4B 5.49% Mar 2026 huggingface.co Official leaderboard
68 qvac/TranscribePsy-250M-EU-sub 8.39% huggingface.co Official leaderboard
#ModelScoreSettingDateLinkSource
1 Voxtral Small 1.84% Jul 2025 huggingface.co Official leaderboard
2 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 1.91% Oct 2026 huggingface.co Official leaderboard
3 Voxtral Mini Transcribe Realtime (2602) 2.23% Feb 2026 huggingface.co Official leaderboard
4 GPT-Transcribe 2.23% Jul 2026 platform.openai.com Official leaderboard
5 Whisper large-v3 2.71% Nov 2023 huggingface.co Official leaderboard
6 Cohere Transcribe 2.74% Mar 2026 huggingface.co Official leaderboard
7 Whisper large-v3-turbo 2.79% Oct 2024 huggingface.co Official leaderboard
8 Muse Voice Transcribe 1.0 3.21% Sep 2026 developer.meta.com Official leaderboard
9 Gemma 4 E4B 5.49% Mar 2026 huggingface.co Official leaderboard

2 models have held number one on Open ASR Leaderboard’s leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI Mistral AI
Open ASR Leaderboard leaderboard 2 leaders

Source: Open ASR Leaderboard’s leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On Open ASR Leaderboard’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Voxtral Small Whisper large-v3 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Voxtral Small

Source: Open ASR Leaderboard’s 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

All 695 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 (68 entries).
  • Pass Models scored 9 models we track, plus 59 other entries on its leaderboard.
  • Pass Newest score The newest is from October 2026, for NVIDIA Nemotron 3.5 Transcribe 30B-A4B.
  • 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 Open ASR Leaderboard: 9 rows for models we track and 59 other entries at the last read.
  • Pass Last read November 20th, 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.

How we check