Alibaba
availableShows if the model has enough results for an index.Qwen3.5 397B
Qwen3.5 397B is a non-reasoning model from Alibaba. 34 benchmarks count toward its score, in 7 categories.
IndexOverall score. 50 is the middle.51.9 ±3.3
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.86/s
Input / 1MUS dollars per 1M input tokens.$0.6
Output / 1MUS dollars per 1M output tokens.$3.6
ContextMaximum tokens in one request.128K
EloLMArena rating and rank.N/A
50 is the middle of the board. The range shows the doubt in the index.
CapabilitiesScore per category. 50 is the middle.
50 is the middleResults
34 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result on the index scale. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| V* | Multimodal | 95.8% | 56.2 | — | — | Z.AI |
| MMLU-Redux | Knowledge | 94.9% | 53.8 | — | — | Qwen |
| Harvard-MIT Mathematics Tournament February 2025 | Math | 94.8% | 53.3 | — | — | Qwen |
| AIME 2026 | Math | 93.3% | 52.9 | — | — | Qwen |
| C-Eval | Knowledge | 93.0% | — | — | — | C-Eval authors |
| Harvard-MIT Mathematics Tournament November 2025 | Math | 92.7% | — | — | — | Qwen |
| Instruction-Following Eval | Instruction | 92.6% | 50.8 | — | — | Jeffrey Zhou et al. |
| MathVision | Multimodal | 88.6% | — | — | — | Qwen |
| Graduate-Level Google-Proof Q&A | Knowledge | 88.4% | 59.9 | — | — | David Rein et al. |
| Harvard-MIT Mathematics Tournament February 2026 | Math | 87.9% | 55.9 | — | — | Qwen |
| Massive Multitask Language Understanding Professional | Knowledge | 87.8% | 58.9 | — | — | Yubo Wang et al. |
| Artificial Analysis GPQA Diamond | Knowledge | 86.1% | 56.8 | — | — | Artificial Analysis |
| VideoMMMU | Multimodal | 84.7% | — | — | — | Qwen |
| MMLU-ProX | Multilingual | 84.7% | — | — | — | MMLU-ProX authors |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 83.9% | 59.0 | — | — | Victor Barres et al. |
| LiveCodeBench v6 | Coding | 83.6% | 53.1 | — | — | LiveCodeBench maintainers |
| MMAnswerBench | Math | 80.9% | — | — | — | Qwen |
| CharXiv Reasoning | Multimodal | 80.8% | 58.1 | — | — | CharXiv authors |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 79.0% | 55.9 | — | — | MMMU-Pro authors |
| Software Engineering Benchmark Verified | Coding | 76.2% | 58.7 | — | — | Carlos E. Jimenez et al. |
| MCP-Tasks | Agentic | 74.2% | — | — | — | Qwen |
| WideResearch | Agentic | 74.0% | 59.3 | — | — | Qwen |
| SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines | Knowledge | 70.4% | 55.8 | — | — | Xiaoxuan Du et al. |
| AI-Needle | Reasoning | 68.7% | — | — | — | Qwen |
| τ³-Bench Tool-Agent-User Evaluation | Agentic | 68.4% | 53.3 | — | — | Sierra Research |
| ScreenSpot Pro | Multimodal | 65.6% | 46.9 | — | — | Kaixin Li et al. |
| Artificial Analysis Long Context Reasoning | Reasoning | 64.3% | 52.7 | — | — | Artificial Analysis |
| LongBench v2 | Reasoning | 63.2% | — | — | — | LongBench v2 authors |
| BrowseComp | Agentic | 62.0% | 52.0 | — | — | OpenAI |
| NOVA-63 | Multilingual | 59.1% | — | — | — | Qwen |
| Claw-Eval | Agentic | 56.8% | 47.7 | — | — | Bowen Ye et al. |
| Artificial Analysis MMMU-Pro | Multimodal | 52.7% | 31.5 | — | — | Artificial Analysis |
| QwenClawBench | Agentic | 51.8% | 51.3 | — | — | Qwen |
| Artificial Analysis IFBench | Instruction | 51.6% | 41.5 | — | — | Artificial Analysis |
| SWE-bench Pro | Coding | 50.9% | 53.2 | — | — | Xiang Deng et al. |
| Gert Labs Composite Game Benchmark | Agentic | 46.8% | 54.0 | — | — | Gert Labs |
| MCP Atlas | Agentic | 46.1% | 43.0 | — | — | OpenAI |
| VITA-Bench | Agentic | 43.7% | 59.2 | — | — | Meituan LongCat Team |
| DeepPlanning | Agentic | 37.6% | — | — | — | DeepPlanning authors |
| Toolathlon | Agentic | 36.3% | 51.8 | — | — | OpenAI |
| Humanity's Last Exam | Knowledge | 28.7% | 53.1 | — | — | Center for AI Safety et al. |
| Artificial Analysis Omniscience Accuracy | Knowledge | 24.5% | 44.2 | — | — | Artificial Analysis |
| Artificial Analysis Intelligence Index | Knowledge | 21.4% | 49.2 | — | — | Artificial Analysis |
| Artificial Analysis Humanity's Last Exam | Knowledge | 19.8% | 46.0 | — | — | Artificial Analysis |
| ResearchClawBench | Agentic | 14.2% | — | — | — | InternScience |
| Critical Physics Tasks | Reasoning | 0.9% | 40.3 | — | — | Artificial Analysis |
34 benchmarks count, from 34 of 46 results. A grey row does not count. Too few models took that benchmark.