- DailyOmni, Qwen3.5-Omni-Flash 81.8
- WorldSense, Qwen3.5-Omni-Flash 57.9
- AVUT, Qwen3.5-Omni-Flash 81.4
- AV-SpeakerBench, Qwen3.5-Omni-Flash 65.2
- VideoMME with audio, Qwen3.5-Omni-Flash 79.3
- QualcommInteractive, Qwen3.5-Omni-Flash 66.3
- Omni-Cloze, Qwen3.5-Omni-Flash 63
- OmniGAIA without thinking prompt, Qwen3.5-Omni-Flash 33.9
- MMAU, Qwen3.5-Omni-Flash 80.4
- MMAR, Qwen3.5-Omni-Flash 74
- MMSU, Qwen3.5-Omni-Flash 72.2
- RUL-MuchoMusic, Qwen3.5-Omni-Flash 60.5
- VoiceBench, Qwen3.5-Omni-Flash 87.8
- SpeechRole, Qwen3.5-Omni-Flash 119.8
- WildSpeech-Bench, Qwen3.5-Omni-Flash 72.2
- Fleurs top60 ASR WER, Qwen3.5-Omni-Flash 10.8
Qwen3.5-Omni-Flash
Qwen3.5-Omni-Flash.
16
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Qwen Mar 2026 Qwen3.5-Omni-Flash
Closed weights Mid-size
No Epoch Capabilities Index score yet
- MMAU 80.4% #4 /15
- WorldSense (omnimodal) 57.8% #4 /12
- OmniGAIA 33.9% #4 /5
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OmniGAIA | 33.9% | #4 of 5 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMAU | 80.4% | #4 of 15 | Lab model card | |
| WorldSense (omnimodal) | 57.8% | #4 of 12 | Official leaderboard | |
| Video-MME | 79.3% | #29 of 51 | Lab model card | |
| VoiceBench | 87.8% | #7 of 12 | Lab model card | |
| MMSU | 72.2% | #4 of 6 | Lab model card | |
| Daily-Omni | 81.8% | #5 of 7 | Lab model card | |
| MMAR | 74% | #3 of 4 | Lab model card | |
| MuChoMusic | 60.5% | #3 of 4 | Lab model card | |
| QIVD | 66.3% | #3 of 4 | Lab model card | |
| FLEURS | 10.8% | #6 of 6 | Lab model card | |
| AV-SpeakerBench | 65.2% | #3 of 3 | Lab model card | |
| AVUT | 81.4% | #3 of 3 | Lab model card | |
| Omni-Cloze | 63% | #3 of 3 | Lab model card | |
| WildSpeech-Bench | 72.2 | #3 of 3 | Lab model card | |
| SpeechRole | 119.8 | #2 of 2 | Lab model card |
Its lab's numbers next to 3 leaderboards
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
| Benchmark | Lab says | Board says | Gap |
|---|---|---|---|
| WorldSense (omnimodal) | 57.9% WorldSense, Qwen3.5-Omni-Flash | 57.8% Leaderboard, modality A+V | Same |
| Omni-Cloze | 63% Omni-Cloze, Qwen3.5-Omni-Flash | 63% Leaderboard | Same |
| OmniGAIA | 33.9% OmniGAIA without thinking prompt, Qwen3.5-Omni-Flash | 33.9% Leaderboard | Same |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.