Qwen3.5-397B-A17B
- Made by
- Alibaba
- Released
- Announced in
- Qwen3.5: Towards Native Multimodal Agentsqwen.ai
- Event
- Qwen3.5
- API record
- /api/v1/models/qwen3-5-397b-a17b
Benchmark results
| Benchmark | Score | Source check |
|---|---|---|
| Humanity's Last ExamNo tools | 28.7% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| Humanity's Last ExamWith tools · search agent; context folding at 256K tokens | 48.3% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| SWE-bench Verified | 76.4% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.CorroboratedThe same number is printed in 1 other lab document: 1 |
| internal harness with bash and file-edit tools; 200K context · reported later, in Qwen3.6-Plus: Towards Real World Agents | 76.2% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| SWE-bench Prointernal harness with bash and file-edit tools; 200K context · reported later, in Qwen3.6-Plus: Towards Real World Agents · Run on the lab's own corrected version of the public set | 50.9% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| GPQA Diamond | 88.4% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| BrowseCompdiscard-all context strategy | 78.6% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.CorroboratedThe same number is printed in 1 other lab document: 1 |
| simple context folding | 69.0% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| Terminal-Bench 2.0 | 52.5% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| OSWorld-Verified | 62.2% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| MMLU-Pro | 87.8% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| HMMT February 2025No tools | 94.8% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| LiveCodeBenchLiveCodeBench v6 | 83.6% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| MCP Atlaspublic set; Gemini 2.5 Pro judge · reported later, in Qwen3.6-Plus: Towards Real World Agents | 74.2% | CheckedA person checked this number against its source on 19 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
Sources
Reported by Alibaba
- Qwen3.5: Towards Native Multimodal Agentsqwen.ai · 16 Feb 2026
Our capture of the source 
Our capture of the source - Qwen3.6-Plus: Towards Real World Agentsqwen.ai · 2 Apr 2026
Our capture of the source