Codesota · Speech · Speech Recognition · AMI-IHMTasks/Speech/Speech Recognition
Speech Recognition · benchmark dataset · 2005 · EN

AMI Meeting Corpus — Individual Headset Microphone.

The AMI Meeting Corpus IHM subset consists of ~100 hours of recorded English meetings captured with individual headset microphones. Long-form spontaneous speech across overlapping speakers makes it a standard stress-test for ASR systems beyond clean read speech.

Paper ↗Submit a result ↵
§ 01 · Leaderboard

Best published scores.

50 results indexed across 1 metric. Shaded row marks current SOTA; ties broken by submission date.

Primary
wer · lower is better
wer· primary
50 rows
#ModelOrgSubmittedPaper / codewer
01Granite Speech 4.1 2BOpenIBMMay 2025Granite-speech: open-source speech-aware LLMs with stron…8.09
02Cohere Transcribe (Mar 2026)OpenCohereMar 2026pwc-dump8.13
03Granite Speech 3.3 2BOpenIBMMay 2025Granite-speech: open-source speech-aware LLMs with stron…8.90
04SYMPHONY-ASR—Jan 2026pwc-dump9.56
05Qwen3-ASR-1.7BOpenAlibabaJan 2026Qwen3-ASR Technical Report · code10.56
06Phi-4 Multimodal InstructOpenMicrosoftMar 2025Phi-4-Mini Technical Report: Compact yet Powerful Multim…11.09
07Parakeet-tdt-0.6b-v2—Apr 2023Efficient Sequence Transduction by Jointly Predicting To… · code11.16
08Moonshine-streaming-small—Jan 2026pwc-dump12.53
09Distil-large-v2—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code14.67
10Distil-large-v3—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code15.16
11Owsm_ctc_v3.1_1B—Jan 2024OWSM v3.1: Better and Faster Open Whisper-Style Speech M… · code15.61
12Niagara-38m-batch.en—Feb 2026pwc-dump15.87
13Parakeet-tdt_ctc-110m—Apr 2023Efficient Sequence Transduction by Jointly Predicting To… · code15.89
14Stt_en_conformer_ctc_large—May 2020Conformer: Convolution-augmented Transformer for Speech … · code15.95
15Distil-medium.en—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code16.12
16Whisper Large v3 TurboOpenOpenAIDec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code16.13
17Distil-small.en—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code16.16
18Parakeet-ctc-0.6b—May 2023Fast Conformer with Linearly Scalable Attention for Effi…16.46
19Whisper-medium.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code16.68
20Whisper Large—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code16.73
21Whisper Large v2OpenOpenAIDec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code16.74
22Lite-whisper-large-v3-turbo-acc—Feb 2025LiteASR: Efficient Automatic Speech Recognition with Low… · code16.97
23Voxtral-Mini-4B-Realtime-2602OpenMistral AIFeb 2026Voxtral Realtime17.07
24VibeVoice-ASR-HF—Jan 2026VIBEVOICE-ASR Technical Report17.20
25Parakeet-rnnt-0.6b—May 2023Fast Conformer with Linearly Scalable Attention for Effi…17.40
26Moonshine-base—Oct 2024Moonshine: Speech Recognition for Live Transcription and… · code17.49
27Whisper-small.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code17.93
28Stt_en_fastconformer_ctc_large—May 2023Fast Conformer with Linearly Scalable Attention for Effi…18.61
29Niagara-19m-batch.en—Feb 2026pwc-dump18.86
30Moonshine-streaming-tiny—Jan 2026pwc-dump19.02
31Stt_en_fastconformer_transducer_large—May 2023Fast Conformer with Linearly Scalable Attention for Effi…19.09
32Asr-conformer-loquacious—Feb 2025pwc-dump19.60
33Lite-whisper-large-v3-fast—Feb 2025LiteASR: Efficient Automatic Speech Recognition with Low… · code19.87
34Stt_en_conformer_ctc_small—May 2020Conformer: Convolution-augmented Transformer for Speech … · code20.43
35Whisper-base.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code21.13
36Moonshine-tiny—Oct 2024Moonshine: Speech Recognition for Live Transcription and… · code22.84
37Whisper-tiny.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code24.24
38Asr-wav2vec2-librispeech—Jun 2021SpeechBrain: A General-Purpose Speech Toolkit · code32.05
39Wav2vec2-large-960h-lv60-self—Jun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code36.77
40Wav2vec2-large-robust-ft-libri-960h—Apr 2021Robust wav2vec 2.0: Analyzing Domain Shift in Self-Super… · code37.75
41Hubert-xlarge-ls960-ft—Jun 2021HuBERT: Self-Supervised Speech Representation Learning b… · code39.11
42Hubert-large-ls960-ft—Jun 2021HuBERT: Self-Supervised Speech Representation Learning b… · code39.72
43Data2vec-audio-large-960h—Feb 2022data2vec: A General Framework for Self-supervised Learni… · code40.51
44Mms-1b-all—May 2023Scaling Speech Technology to 1,000+ Languages · code42.02
45Wav2vec2-conformer-rel-pos-large-960h-ft—Oct 2020fairseq S2T: Fast Speech-to-Text Modeling with fairseq · code42.39
46Wav2vec2-conformer-rope-large-960h-ft—Oct 2020fairseq S2T: Fast Speech-to-Text Modeling with fairseq · code42.47
47wav2vec 2.0 Large (960h)OpenMeta AIJun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code42.66
48Wav2vec2-base-960h—Jun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code45.56
49Data2vec-audio-base-960h—Feb 2022data2vec: A General Framework for Self-supervised Learni… · code47.27
50Mms-1b-fl102—May 2023Scaling Speech Technology to 1,000+ Languages · code86.78
Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.
§ 03 · Progress

4 steps
of state of the art.

Each row below marks a model that broke the previous record on wer. Intermediate submissions are kept in the leaderboard above; only SOTA-setting entries are re-listed here.

Lower scores win. Each subsequent entry improved upon the previous best.

SOTA line · wer
  1. May 16, 2020Stt_en_conformer_ctc_large15.95
  2. Apr 13, 2023Parakeet-tdt-0.6b-v211.16
  3. Mar 3, 2025Phi-4 Multimodal InstructMicrosoft11.09
  4. May 13, 2025Granite Speech 4.1 2BIBM8.09
Fig 3 · SOTA-setting models only. 4 entries span May 2020 → May 2025.
§ 04 · Literature

20 papers
tied to this benchmark.

Every paper below corresponds to at least one row in the leaderboard above. Click through for the arXiv preprint and, when available, the reference implementation.

§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies