Qwen in-house voice-chat benchmark
Models answer spoken versions of text benchmarks such as MMLU, CEval, IFEval, GSM8K, Math23K and Math401. Each part keeps the source benchmark's scoring.
In-house Qwen evaluation. The tasks are not public. About 90% of the text instructions suit speech and are used.
- Made by
- Qwen (Alibaba)
- Released
- March 2025
- Kind
- Speech recognition and translation
- Level
- School
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- Kept private
- Last in a system card
- March 2025
- Last model added
- March 2025
1
System card reports it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Qwen in-house voice-chat benchmark yet.
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
One system card reports it
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Mar 26, 2025 |  Qwen2.5-Omni-7B technical report | In-house voice-chat benchmark with speech instructions: MMLU* 65.6, CEval* 61.1, IFEval* 41.7, GSM8K* 85.4, Math23K* 87.1, Math401* 62.2, no overall score (Table 4) | Capabilities |
More 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%