Alibaba
availableShows if the model has enough results for an index.Qwen3.8 Max
Qwen3.8 Max is a reasoning model from Alibaba. 48 benchmarks count toward its score, in 7 categories.
IndexOverall score out of 100.68.4 ±2.8
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.39/s
Input / 1MUS dollars per 1M input tokens.N/A
Output / 1MUS dollars per 1M output tokens.N/A
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.N/A
The index is a score out of 100. The ± range shows how much it can change.
CapabilitiesScore per category, out of 100.
Out of 100Results
48 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result as a score out of 100. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | Math | 99.4% | 67.1 | xhigh effort | — | Epoch AI |
| MathVision with Python | Multimodal | 97.7% | — | — | — | Moonshot AI / MathVision authors |
| MathVision | Multimodal | 95.2% | — | — | — | Qwen |
| GPQA Diamond | Knowledge | 93.7% | 64.7 | — | 1 Sept 2026 | Vals AI |
| CharXiv Reasoning | Multimodal | 93.5% | 72.6 | — | — | CharXiv authors |
| PaperBench | Coding | 93.0% | — | — | — | Qwen Team |
| MRCRv2 | Reasoning | 92.9% | — | — | — | OpenAI |
| GPQA diamond | Knowledge | 92.7% | 63.8 | xhigh effort | — | Epoch AI |
| Graduate-Level Google-Proof Q&A | Knowledge | 92.6% | 63.7 | — | — | David Rein et al. |
| GPQA Diamond | Knowledge | 92.6% | 63.7 | — | — | David Rein et al. |
| OmniDocBench 1.5 | Multimodal | 92.1% | — | — | — | OpenAI |
| LiveBench Mathematics | Math | 91.3% | 69.9 | — | 25 Jun 2026 | LiveBench |
| BabyVision with Python | Multimodal | 91.3% | — | — | — | Moonshot AI |
| MLVU mean average | Multimodal | 90.8% | — | — | — | Qwen |
| Video-MME with subtitle | Multimodal | 90.4% | — | — | — | Qwen |
| VideoMMMU | Multimodal | 88.7% | — | — | — | Qwen |
| MMLU Pro | Knowledge | 88.6% | 60.1 | — | 1 Sept 2026 | Vals AI |
| CharXiv Reasoning without tools | Multimodal | 88.4% | — | — | — | CharXiv authors |
| LiveBench Reasoning | Reasoning | 88.2% | 78.5 | — | 25 Jun 2026 | LiveBench |
| MMMU Pro | Multimodal | 88.0% | 70.6 | — | 1 Sept 2026 | Vals AI |
| RealWorldQA | Multimodal | 88.0% | 62.2 | — | — | Qwen |
| LiveCodeBench | Coding | 87.9% | 64.2 | — | 1 Sept 2026 | Vals AI |
| Terminal-Bench 2.1 (provider run) | Agentic | 86.6% | 75.2 | — | — | DeepSeek-AI |
| Terminal-Bench 2.1 (provider run) | Agentic | 86.6% | 75.2 | — | — | DeepSeek-AI |
| OSWorld-Verified | Agentic | 86.1% | 73.9 | — | — | Tianbao Xie et al. |
| SWE-bench | Coding | 85.6% | 66.2 | — | 1 Sept 2026 | Vals AI |
| AndroidWorld | Agentic | 85.3% | — | — | — | Z.AI |
| ScreenSpot Pro | Multimodal | 84.5% | 66.5 | — | — | Kaixin Li et al. |
| Instruction Following Benchmark | Instruction | 82.8% | 62.4 | — | — | Benchmark authors |
| Multimodal Multi-disciplinary Video Understanding | Multimodal | 82.4% | — | — | — | MMVU benchmark maintainers |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 82.3% | 61.3 | — | — | MMMU-Pro authors |
| BabyVision | Multimodal | 82.0% | — | — | — | Meta AI |
| WideResearch | Agentic | 81.9% | 67.9 | — | — | Qwen |
| LVBench | Multimodal | 81.8% | — | — | — | Qwen Team |
| VulcanBench v3 | Coding | 81.2% | 63.3 | — | — | VulcanBench contributors |
| MedXpertQA Multimodal | Multimodal | 80.4% | — | — | — | Meta AI |
| LiveBench Language | Knowledge | 79.7% | 66.7 | — | 25 Jun 2026 | LiveBench |
| CC-OCR | Multimodal | 79.6% | — | — | — | Qwen |
| LiveBench Data Analysis | Reasoning | 78.4% | 64.9 | — | 25 Jun 2026 | LiveBench |
| MobileWorld | Agentic | 77.8% | — | — | — | Qwen Team |
| ERQA | Multimodal | 77.8% | 71.7 | — | — | Qwen |
| SimpleVQA | Multimodal | 75.0% | 68.2 | — | — | Z.AI |
| CoWorkBench | Agentic | 74.8% | — | — | — | Qwen Team |
| FrontierMath-Tiers-1-3-v2-Private | Math | 74.7% | 73.5 | xhigh effort | — | Epoch AI |
| OCRBench V2 | Multimodal | 74.2% | — | — | — | OCRBench authors |
| LiveBench Instruction Following | Instruction | 74.1% | 76.9 | — | 25 Jun 2026 | LiveBench |
| FrontierSWE | Coding | 73.5% | — | — | — | Evan Chu et al. |
| IOI v1 | Coding | 73.0% | 81.5 | — | 9 Aug 2026 | Vals AI |
| LiveBench Coding | Coding | 72.9% | 58.7 | — | 25 Jun 2026 | LiveBench |
| Toolathlon-Verified | Agentic | 72.5% | 71.0 | — | — | Moonshot AI |
| Vision2Web | Multimodal | 69.0% | — | — | — | Z.AI |
| IOI | Coding | 68.9% | 75.7 | — | 21 Sept 2026 | Vals AI |
| SWE-bench Pro | Coding | 67.7% | 69.5 | — | — | Xiang Deng et al. |
| Terminal-Bench 2.1 | Agentic | 67.4% | 63.8 | — | 21 Sept 2026 | Vals AI |
| WebArena-Verified Browser Agent Benchmark | Agentic | 66.8% | — | — | — | Amine El Hattami et al. |
| LongBench v2 | Reasoning | 66.3% | — | — | — | LongBench v2 authors |
| Vibe Code Bench v1.1 | Coding | 64.7% | 69.1 | OpenHands | 21 Sept 2026 | Vals AI |
| LiveBench Agentic Coding | Agentic | 64.6% | 77.5 | — | 25 Jun 2026 | LiveBench |
| PerceptionBench (Internal) | Multimodal | 63.5% | — | — | — | Moonshot AI |
| OpenHarmony Bench v1.0 | Coding | 60.8% | 70.0 | — | — | OpenHarmony Bench authors |
| ProofBench v1.1 | Math | 58.0% | 72.9 | — | 21 Sept 2026 | Vals AI |
| DeepSWE | Agentic | 56.6% | 64.9 | — | — | Datacurve AI |
| Humanity's Last Exam with tools | Agentic | 56.2% | 69.0 | — | — | DeepSeek-AI |
| NL2Repo | Coding | 55.9% | 68.9 | — | — | MiniMax |
| τ²-bench Banking | Agentic | 55.2% | 38.4 | xhigh effort · Sierra | 4 Aug 2026 | Sierra Research |
| JobBench | Agentic | 53.4% | 70.8 | — | — | Yuetai Li et al. |
| Agents' Last Exam | Agentic | 52.4% | 90.5 | — | — | DeepSeek-AI |
| ZeroBench_main with Python | Multimodal | 49.0% | — | — | — | Moonshot AI / ZeroBench authors |
| FrontierMath-Tier-4-v2-Private | Math | 46.3% | 70.2 | xhigh effort | — | Epoch AI |
| SimpleQA Verified | Knowledge | 45.8% | 64.0 | xhigh effort | — | Epoch AI |
| Humanity's Last Exam | Knowledge | 43.6% | 65.7 | — | — | Center for AI Safety et al. |
| Humanity's Last Exam without tools | Knowledge | 43.6% | 65.7 | — | — | OpenAI |
| SkillsBench | Coding | 42.0% | 58.6 | OpenHands | 11 Sept 2026 | Vals AI |
| MLS-Bench Lite | Coding | 41.0% | — | — | — | MLS-Bench |
| Mystery Game Puzzles | Reasoning | 38.0% | 74.3 | xhigh effort | — | Epoch AI |
| Chess Puzzles | Reasoning | 29.0% | 61.0 | xhigh effort | — | Epoch AI |
| AutomationBench | Agentic | 27.3% | 60.7 | — | — | Moonshot AI |
| Terminal-Bench 4.0 | Agentic | 24.7% | 76.9 | — | 21 Sept 2026 | Vals AI |
| ZeroBench | Multimodal | 24.0% | — | — | — | Meta AI |
| Code Migration | Coding | 24.0% | 60.2 | — | 21 Sept 2026 | Vals AI |
| OSWorld 2.0 | Agentic | 19.4% | 65.2 | — | — | Mengqi Yuan et al. |
| FrontierSWE v2 | Coding | 15.8% | 63.2 | — | — | Proximal |
| Vibe Code Bench 1-100 | Coding | 12.8% | 65.9 | OpenHands | 16 Sept 2026 | Vals AI |
| ProgramBench | Coding | 0.0% | — | — | 21 Sept 2026 | Vals AI |
48 benchmarks count, from 56 of 84 results. A grey row does not count. Too few models took that benchmark.