Codesota · Speech · Speech Recognition · GigaSpeechTasks/Speech/Speech Recognition
Speech Recognition · benchmark dataset · 2021 · EN

GigaSpeech.

GigaSpeech is a 10,000-hour English ASR corpus pulled from audiobooks, podcasts, and YouTube. Released by SpeechColab and widely used as a high-volume training+evaluation set covering diverse speaking styles and noise conditions.

Paper ↗Submit a result ↵
§ 01 · Leaderboard

Best published scores.

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

Primary
wer · lower is better
wer· primary
47 rows
#ModelOrgSubmittedPaper / codewer
01Qwen3-ASR-1.7BOpenAlibabaJan 2026Qwen3-ASR Technical Report · code8.74
02Phi-4 Multimodal InstructOpenMicrosoftMar 2025Phi-4-Mini Technical Report: Compact yet Powerful Multim…9.33
03Canary-Qwen-2.5BOpenNVIDIAMar 2025Training and Inference Efficiency of Encoder-Decoder Spe…9.43
04VibeVoice-ASR-HF—Jan 2026VIBEVOICE-ASR Technical Report9.67
05Granite Speech 4.1 2BOpenIBMMay 2025Granite-speech: open-source speech-aware LLMs with stron…9.80
06SYMPHONY-ASR—Jan 2026pwc-dump9.96
07Parakeet-rnnt-0.6b—May 2023Fast Conformer with Linearly Scalable Attention for Effi…10.01
08Distil-large-v3—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code10.08
09Whisper Large v3 TurboOpenOpenAIDec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code10.14
10Lite-whisper-large-v3-turbo-acc—Feb 2025LiteASR: Efficient Automatic Speech Recognition with Low… · code10.21
11Audio Flamingo 3—Jul 2025Audio Flamingo 3: Advancing Audio Intelligence with Full… · code10.27
12Distil-large-v2—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code10.32
13Voxtral-Mini-4B-Realtime-2602OpenMistral AIFeb 2026Voxtral Realtime10.38
14Parakeet-ctc-0.6b—May 2023Fast Conformer with Linearly Scalable Attention for Effi…10.39
15Parakeet-tdt_ctc-110m—Apr 2023Efficient Sequence Transduction by Jointly Predicting To… · code10.52
16Lite-whisper-large-v3-fast—Feb 2025LiteASR: Efficient Automatic Speech Recognition with Low… · code10.66
17Whisper Large v2OpenOpenAIDec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code10.67
18Granite Speech 3.3 2BOpenIBMMay 2025Granite-speech: open-source speech-aware LLMs with stron…10.69
19Whisper Large—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code10.76
20Distil-small.en—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code10.87
21Whisper-medium.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code11.03
22Distil-medium.en—Nov 2023Distil-Whisper: Robust Knowledge Distillation via Large-… · code11.30
23Whisper-small.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code11.35
24Asr-conformer-loquacious—Feb 2025pwc-dump11.36
25Stt_en_conformer_ctc_large—May 2020Conformer: Convolution-augmented Transformer for Speech … · code11.60
26Moonshine-base—Oct 2024Moonshine: Speech Recognition for Live Transcription and… · code12.08
27Stt_en_fastconformer_ctc_large—May 2023Fast Conformer with Linearly Scalable Attention for Effi…12.17
28Stt_en_fastconformer_transducer_large—May 2023Fast Conformer with Linearly Scalable Attention for Effi…12.31
29Niagara-38m-batch.en—Feb 2026pwc-dump12.56
30Whisper-base.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code12.83
31Moonshine-streaming-tiny—Jan 2026pwc-dump13.90
32Whisper-tiny.en—Dec 2022Robust Speech Recognition via Large-Scale Weak Supervisi… · code14.08
33Moonshine-tiny—Oct 2024Moonshine: Speech Recognition for Live Transcription and… · code14.21
34Stt_en_conformer_ctc_small—May 2020Conformer: Convolution-augmented Transformer for Speech … · code14.46
35Asr-wav2vec2-librispeech—Jun 2021SpeechBrain: A General-Purpose Speech Toolkit · code16.92
36Wav2vec2-large-960h-lv60-self—Jun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code23.94
37Hubert-xlarge-ls960-ft—Jun 2021HuBERT: Self-Supervised Speech Representation Learning b… · code24.74
38Data2vec-audio-large-960h—Feb 2022data2vec: A General Framework for Self-supervised Learni… · code24.80
39Wav2vec2-conformer-rel-pos-large-960h-ft—Oct 2020fairseq S2T: Fast Speech-to-Text Modeling with fairseq · code24.96
40Wav2vec2-conformer-rope-large-960h-ft—Oct 2020fairseq S2T: Fast Speech-to-Text Modeling with fairseq · code25
41Hubert-large-ls960-ft—Jun 2021HuBERT: Self-Supervised Speech Representation Learning b… · code25.01
42Wav2vec2-large-robust-ft-libri-960h—Apr 2021Robust wav2vec 2.0: Analyzing Domain Shift in Self-Super… · code25.12
43Mms-1b-all—May 2023Scaling Speech Technology to 1,000+ Languages · code26.44
44wav2vec 2.0 Large (960h)OpenMeta AIJun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code27.74
45Data2vec-audio-base-960h—Feb 2022data2vec: A General Framework for Self-supervised Learni… · code29.78
46Wav2vec2-base-960h—Jun 2020wav2vec 2.0: A Framework for Self-Supervised Learning of… · code30.85
47Mms-1b-fl102—May 2023Scaling Speech Technology to 1,000+ Languages · code42.42
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

5 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_large11.60
  2. Dec 6, 2022Whisper Large v3 TurboOpenAI10.14
  3. May 8, 2023Parakeet-rnnt-0.6b10.01
  4. Mar 3, 2025Phi-4 Multimodal InstructMicrosoft9.33
  5. Jan 29, 2026Qwen3-ASR-1.7BAlibaba8.74
Fig 3 · SOTA-setting models only. 5 entries span May 2020 → Jan 2026.
§ 04 · Literature

21 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