Benchmark Almanac

MU-Bench

Chirp-3 14.1% tops its leaderboard gpt-4o-mini-transcribe 27.3% leads the models we track Not in model cards

Speech-to-text providers transcribe 4,270 caller utterances from real 8 kHz customer service phone calls in five locales, scored against human word-level transcripts.

Locales are en-US, es-MX, tr-TR, vi-VN and zh-CN. Data is CC BY-NC 4.0.

Made by Sierra Released Apr 2026 Speech recognition and translation All tasks public Last model added Mar 2025 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Short answer A number, word or short string matched to a key.
sierra-research/mu-bench 18 stars Paper Dataset Leaderboard
8
Entries on its own leaderboard, 1 model we track

Chirp-3 tops the leaderboard with 14.1%

Best score per entry from the benchmark's own leaderboard, read on 2 Nov 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 gpt-4o-mini-transcribe, at 27.3%.

  1. 1 Chirp-3 14.1%
  2. 2 Gemini-3.5-Transcribe 14.9%
  3. 3 Scribe-v2 21.2%
  4. 4 Qwen3-ASR-1.7B 23.6%
  5. 5 Azure-Speech-v1 24.6%
  6. 6 gpt-4o-mini-transcribe 27.3%
  7. 7 Grok-STT 32.1%
  8. 8 Nova-3 32.6%
#EntryScoreSettingDateSource
1 Chirp-3 14.1% Official leaderboard
2 Gemini-3.5-Transcribe 14.9% Official leaderboard
3 Scribe-v2 21.2% Official leaderboard
4 Qwen3-ASR-1.7B 23.6% Baseten Official leaderboard
5 Azure-Speech-v1 24.6% Official leaderboard
6 gpt-4o-mini-transcribe 27.3% Mar 2025 Official leaderboard
7 Grok-STT 32.1% Official leaderboard
8 Nova-3 32.6% Official leaderboard

Scores over time

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

On the benchmark’s own leaderboard
0% 20% 40% 60% 80% 100% 2025 gpt-4o-mini-transcribe 0% 20% 40% 60% 80% 100% 2025 gpt-4o-mini-transcribe

No lab model card reports it yet

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.

More in Audio and video

All 168 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%
  • 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%
  • TVBench In use Models answer multiple-choice questions on 10 temporal video tasks, such as action counting and ordering, that static frames cannot solve. Scored by accuracy. Qwen 3.8 Max 81.9%