MiMo-V2.6-Pro vs Qwen3.8 Max
MiMo-V2.6-Pro 
Qwen3.8 Max Verdict
- MiMo-V2.6-Pro and Qwen3.8 Max (max) are level on quality (64.8 vs 64.2).
- MiMo-V2.6-Pro is stronger in composite, knowledge, long context, maths, reasoning.
- Qwen3.8 Max (max) is stronger in agents & tools, coding, human preference, multimodal.
- MiMo-V2.6-Pro is 5.5× cheaper ($0.548 vs $3.00 per 1M blended).
| Metric | MiMo-V2.6-Pro | Qwen3.8 Max (max) |
|---|---|---|
| BenchLeader Index | 64.8 | 64.2 |
| Agents & tools score | 56.5 | 58.2 |
| Coding score | 62.0 | 64.6 |
| Composite score | 82.1 | 70.7 |
| Human preference score | 67.7 | 68.0 |
| Knowledge score | 65.5 | 62.7 |
| Long context score | 68.5 | 65.4 |
| Maths score | 66.1 | 65.1 |
| Multimodal score | 60.1 | 66.2 |
| Reasoning score | 79.6 | 72.2 |
| Blended price $/M | $0.548 | $3.00 |
| Output speed | 43 tok/s | 37 tok/s |
| Time to first answer | 51.1 s | 58.9 s |
| Context window | 1.0M | 1M |
| Terminal-Bench | – | 27.0% |
| SciCode | 60.9% | 53.2% |
| APEX-Agents | 59.5% | 63.3% |
| ProofBench | 70.0% | 58.0% |
| Epoch Capabilities Index | – | 156.4 |
| LMArena Text | 1480 | 1483 |
| LMArena Hard Prompts | 1506 | 1504 |
| LMArena Coding | 1539 | 1524 |
| LMArena WebDev | 1629 | 1672 |
| LMArena Vision | 1264 | 1314 |
| LMArena Agent | 3.5 | 2.3 |
| LiveBench | – | 78.5% |
| LiveBench Reasoning | – | 88.2% |
| LiveBench Coding | – | 72.9% |
| LiveBench Agentic Coding | – | 64.7% |
| LiveBench Mathematics | – | 91.3% |
| LiveBench Data Analysis | – | 78.4% |
| LiveBench Language | – | 79.7% |
| LiveBench Instruction Following | – | 74.1% |
| AA Intelligence Index v4.3.2 | 46.3 | 45.4 |
| AA-LCR | 86.3% | 80.3% |
| MMMU-Pro | – | 82.8% |
| AA-Omniscience | 8.4 | 12.0 |
| GPQA Diamond (AA) | – | 92.8% |
| Humanity's Last Exam (AA) | 49.4% | 43.1% |
| SciCode (AA) | 60.9% | 53.2% |
| LiveCodeBench | – | 87.8% |
| MMLU-Pro | – | 88.6% |
| IOI | 39.3% | 68.9% |
| LegalBench | – | 83.6% |
| CorpFin | – | 65.8% |
| TaxEval | – | 75.5% |
| Terminal-Bench 2.1 (Vals) | 67.8% | 67.4% |
| SWE-bench (Vals) | – | 85.6% |
| GPQA Diamond (Vals) | – | 93.7% |
| Vals Index | 55.2 | 48.3 |
| CritPt | 26.6% | 20.0% |
| GDPval-AA v2.1 | 59.2% | 58.6% |
| τ³-Banking (AA) | – | 51.3% |
| ITBench SRE (AA) | – | 40.3% |
| Analyst Agent (AA) | – | 45.0% |
| APEX-Agents (AA) | – | 42.4% |
| Code Migration | 43.0% | 24.0% |
| CyberBench | 72.9% | 28.6% |
| Excel Modeling Benchmark | 62.9% | 60.1% |
| Finance Agent v2 | 57.3% | 50.6% |
| Harvey's Legal Agent Benchmark | 10.8% | 10.4% |
| Legal Research Bench | 47.1% | 47.6% |
| MedCode | 45.0% | 40.7% |
| MedScribe | 88.3% | 85.0% |
| MMMU-Pro (Vals) | – | 88.0% |
| MortgageTax | – | 64.0% |
| MysteryMechanism | 15.3% | 23.9% |
| ProgramBench | 0.5% | 0.0% |
| Public Benefits Bench | 68.9% | 67.1% |
| SAGE | 45.0% | 51.3% |
| SkillsBench | – | 42.0% |
| SREBench | 3.0% | – |
| Tax Agent Bench | 64.9% | 66.0% |
| Terminal-Bench 4.0 (Vals) | 31.3% | 34.3% |
| Terminal-Bench Science | 2.9% | 1.4% |
| Vals Multimodal Index | – | 65.4% |
| Vibe Code Bench 1-100 | – | 12.8% |
| Vibe Code Bench v1.1 | 85.2% | 64.7% |
| LMArena Maths | 1483 | 1497 |
| LMArena Creative Writing | 1448 | 1470 |
| LMArena Instruction Following | 1474 | 1474 |
| LMArena Multi-turn | 1456 | 1492 |
| LMArena Longer Queries | 1495 | 1492 |
| ALE-Bench | 1158.1 | – |
| Terminal-Bench 4.0 (AA) | 34.9% | 38.9% |
| Terminal-Bench 2.1 (AA) | – | 88.8% |
| AutomationBench | 58.6% | 56.2% |
| GDP.pdf | 19.2% | 22.8% |
| MLCR | 18.3% | 20.0% |
| EnterpriseOps-Gym | – | 47.6% |
| AA-Omniscience: accuracy | 34.9% | 31.9% |
| AA-Omniscience: non-hallucination | 59.4% | 71.2% |
| AA-Briefcase v1.1 | 1516 | 1617 |
| AA Openness Index | 52.8 | – |
Data as of 2026-10-11. Best configuration of each model; every score links to its source on the model pages.
MiMo-V2.6-Pro vs Qwen3.8 Max: questions
- Is MiMo-V2.6-Pro better than Qwen3.8 Max?
- MiMo-V2.6-Pro and Qwen3.8 Max (max) are level on quality (64.8 vs 64.2). The BenchLeader Index combines every independent quality benchmark; MiMo-V2.6-Pro is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is MiMo-V2.6-Pro better than Qwen3.8 Max for coding?
- Qwen3.8 Max scores higher in coding (65 vs 62 on the category index, where 50 is average).
- Is MiMo-V2.6-Pro better than Qwen3.8 Max for agentic tasks?
- Qwen3.8 Max scores higher in agentic tasks (58 vs 57 on the category index, where 50 is average).
- Which is cheaper, MiMo-V2.6-Pro or Qwen3.8 Max?
- MiMo-V2.6-Pro is cheaper: $0.548 against $3.00 per million tokens, blended at three input tokens per output token.
- Which is faster, MiMo-V2.6-Pro or Qwen3.8 Max?
- MiMo-V2.6-Pro streams faster: 43 against 37 output tokens per second.
- Which has the larger context window?
- MiMo-V2.6-Pro accepts more context: 1.0M against 1M tokens.