Benchmark Almanac

GigaSpeech

Transducer: Zipformer encoder + Embedding decoder 10.4% tops its leaderboard
Not in model cards

Models transcribe English speech from audiobooks, podcasts, and online videos. The test compares transcripts using word error rate.

Made by
GigaSpeech team
Released
June 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
In views
Audio and speech
9
Entries on its own leaderboard, 0 models we track

Transducer: Zipformer encoder + Embedding decoder tops the leaderboard with 10.4%

Best score per entry from the benchmark's own 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.

  1. 1 Transducer: Zipformer encoder + Embedding decoder GigaSpeech v1.0.0 XL 10.4%
  2. 2 Pruned Stateless RNN-T GigaSpeech v1.0.0 XL 10.5%
  3. 3 Conformer CTC + ngram & attention rescoring GigaSpeech v1.0.0 XL 10.6%
  4. 4 Joint CTC/AED(U2++) GigaSpeech v1.0.0 XL 10.6%
  5. 5 Conformer/Transformer-AED GigaSpeech v1.0.0 XL 10.8%
  6. 6 Transformer-AED GigaSpeech v1.0.0 XL 11.6%
  7. 7 RNN-T GigaSpeech v1.0.0 XL 12.3%
  8. 8 Transformer-AED + RNNLM GigaSpeech v1.0.0 XL 12.7%
  9. 9 Chain + RNNLM GigaSpeech v1.0.0 XL 14.8%

Source: Official GigaSpeech leaderboard

#EntryScoreSettingSubmitted byDateLinkSource
1 Transducer: Zipformer encoder + Embedding decoder 10.4% GigaSpeech v1.0.0 XL, Icefall Johns Hopkins University huggingface.co Official leaderboard
2 Pruned Stateless RNN-T 10.5% GigaSpeech v1.0.0 XL, Icefall Johns Hopkins University huggingface.co Official leaderboard
3 Conformer CTC + ngram & attention rescoring 10.6% GigaSpeech v1.0.0 XL, Icefall Johns Hopkins University huggingface.co Official leaderboard
4 Joint CTC/AED(U2++) 10.6% GigaSpeech v1.0.0 XL, Wenet Mobvoi mobvoi-speech-public.ufile.ucloud.cn Official leaderboard
5 Conformer/Transformer-AED 10.8% GigaSpeech v1.0.0 XL, Espnet zenodo.org Official leaderboard
6 Transformer-AED 11.6% GigaSpeech v1.0.0 XL, NeurST ByteDance AI Lab lf3-nlp-opensource.bytetos.com Official leaderboard
7 RNN-T 12.3% GigaSpeech v1.0.0 XL, Pika Official leaderboard
8 Transformer-AED + RNNLM 12.7% GigaSpeech v1.0.0 XL, Athena drive.google.com Official leaderboard
9 Chain + RNNLM 14.8% GigaSpeech v1.0.0 XL, Kaldi Official 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 9 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 9 entries at the last read, none a model we track.
  • 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