Kimi K3 vs Qwen3.8 Max
Kimi K3 
Qwen3.8 Max Verdict
- Kimi K3 (max) leads on quality: 65.5 vs 64.2.
- Kimi K3 (max) is stronger in agents & tools, coding, composite, human preference, knowledge, long context.
- Qwen3.8 Max (max) is stronger in maths, multimodal, reasoning.
- Qwen3.8 Max (max) is 2.0× cheaper ($3.00 vs $6.00 per 1M blended).
| Metric | Kimi K3 (max) | Qwen3.8 Max (max) |
|---|---|---|
| BenchLeader Index | 65.5 | 64.2 |
| Agents & tools score | 66.2 | 58.2 |
| Coding score | 64.7 | 64.6 |
| Composite score | 79.0 | 70.7 |
| Human preference score | 68.6 | 68.0 |
| Knowledge score | 63.2 | 62.7 |
| Long context score | 69.7 | 65.4 |
| Maths score | 63.8 | 65.1 |
| Multimodal score | 63.7 | 66.2 |
| Reasoning score | 66.6 | 72.2 |
| Blended price $/M | $6.00 | $3.00 |
| Output speed | 41 tok/s | 37 tok/s |
| Time to first answer | 52.2 s | 58.9 s |
| Context window | 1.0M | 1M |
| GPQA Diamond | 93.1% | – |
| FrontierMath Tiers 1–3 | 72.2% | – |
| FrontierMath Tier 4 | 39.0% | – |
| OTIS Mock AIME | 97.2% | – |
| SimpleQA Verified | 50.6% | – |
| Terminal-Bench | – | 27.0% |
| SimpleBench | 60.7% | – |
| SciCode | 59.5% | 53.2% |
| WeirdML | 82.6% | – |
| APEX-Agents | – | 63.3% |
| ProofBench | – | 58.0% |
| Epoch Capabilities Index | – | 156.4 |
| LMArena Text | 1488 | 1483 |
| LMArena Hard Prompts | 1517 | 1504 |
| LMArena Coding | 1540 | 1524 |
| LMArena WebDev | 1654 | 1672 |
| LMArena Vision | – | 1314 |
| LMArena Agent | 3.8 | 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 | 43.6 | 45.4 |
| AA-LCR | 88.7% | 80.3% |
| MMMU-Pro | 80.5% | 82.8% |
| AA-Omniscience | 19.7 | 12.0 |
| GPQA Diamond (AA) | 93.5% | 92.8% |
| Humanity's Last Exam (AA) | 46.9% | 43.1% |
| SciCode (AA) | 59.5% | 53.2% |
| LiveCodeBench | 87.2% | 87.8% |
| MMLU-Pro | 88.0% | 88.6% |
| IOI | 48.9% | 68.9% |
| LegalBench | – | 83.6% |
| CorpFin | – | 65.8% |
| TaxEval | – | 75.5% |
| Terminal-Bench 2.1 (Vals) | – | 67.4% |
| SWE-bench (Vals) | – | 85.6% |
| GPQA Diamond (Vals) | – | 93.7% |
| Vals Index | – | 48.3 |
| MCP Atlas | 82.3% | – |
| ARC-AGI-1 | 94.5% | – |
| ARC-AGI-2 | 60.4% | – |
| CritPt | 23.4% | 20.0% |
| GDPval-AA v2.1 | 51.7% | 58.6% |
| τ³-Banking (AA) | 46.0% | 51.3% |
| ITBench SRE (AA) | 47.7% | 40.3% |
| Analyst Agent (AA) | 38.8% | 45.0% |
| APEX-Agents (AA) | 41.3% | 42.4% |
| BioMysteryBench | 71.5% | – |
| Code Migration | 16.1% | 24.0% |
| CyberBench | 75.2% | 28.6% |
| Excel Modeling Benchmark | 66.4% | 60.1% |
| Finance Agent v2 | – | 50.6% |
| Harvey's Legal Agent Benchmark | 10.8% | 10.4% |
| Legal Research Bench | – | 47.6% |
| MedCode | – | 40.7% |
| MedScribe | – | 85.0% |
| MMMU-Pro (Vals) | – | 88.0% |
| MortgageTax | – | 64.0% |
| MysteryMechanism | – | 23.9% |
| ProgramBench | – | 0.0% |
| Public Benefits Bench | 68.3% | 67.1% |
| SAGE | – | 51.3% |
| SkillsBench | – | 42.0% |
| Tax Agent Bench | 68.7% | 66.0% |
| Terminal-Bench 4.0 (Vals) | 19.7% | 34.3% |
| Terminal-Bench Science | 2.9% | 1.4% |
| Time Horizon Index: KSP | 10.5% | – |
| Vals Multimodal Index | – | 65.4% |
| Vibe Code Bench 1-100 | 18.2% | 12.8% |
| Vibe Code Bench v1.1 | – | 64.7% |
| FORTRESS | 26.6% | – |
| LMArena Maths | 1498 | 1497 |
| LMArena Creative Writing | 1460 | 1470 |
| LMArena Instruction Following | 1488 | 1474 |
| LMArena Multi-turn | 1500 | 1492 |
| LMArena Longer Queries | 1504 | 1492 |
| Chess Puzzles | 39.0% | – |
| Mystery Game Puzzles | 26.0% | – |
| Surface Evolver Bench | 93.0% | – |
| DeepSWE v1.1 | 68.5% | – |
| LMCA | 52.7% | – |
| DTBench | 91.2% | – |
| ForecastBench | 61.1% | – |
| ALE-Bench | 1524.5 | – |
| GDP.pdf | 19.0% | – |
| FrontierSWE | 25.9% | – |
| Terminal-Bench 4.0 (AA) | 12.6% | 38.9% |
| Terminal-Bench 2.1 (AA) | 85.0% | 88.8% |
| AutomationBench | 58.3% | 56.2% |
| GDP.pdf | 22.0% | 22.8% |
| MLCR | 38.3% | 20.0% |
| Harvey LAB | 5.3% | – |
| EnterpriseOps-Gym | 45.3% | 47.6% |
| AA-Omniscience: accuracy | 47.6% | 31.9% |
| AA-Omniscience: non-hallucination | 46.8% | 71.2% |
| AA-Briefcase v1.1 | 1501 | 1617 |
| AA Openness Index | 38.9 | – |
Data as of 2026-10-11. Best configuration of each model; every score links to its source on the model pages.
Kimi K3 vs Qwen3.8 Max: questions
- Is Kimi K3 better than Qwen3.8 Max?
- Kimi K3 (max) leads on quality: 65.5 vs 64.2. The BenchLeader Index combines every independent quality benchmark; Kimi K3 (max) is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is Kimi K3 better than Qwen3.8 Max for coding?
- Kimi K3 scores higher in coding (65 vs 65 on the category index, where 50 is average).
- Is Kimi K3 better than Qwen3.8 Max for agentic tasks?
- Kimi K3 scores higher in agentic tasks (66 vs 58 on the category index, where 50 is average).
- Which is cheaper, Kimi K3 or Qwen3.8 Max?
- Qwen3.8 Max is cheaper: $3.00 against $6.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Kimi K3 or Qwen3.8 Max?
- Kimi K3 streams faster: 41 against 37 output tokens per second.
- Which has the larger context window?
- Kimi K3 accepts more context: 1.0M against 1M tokens.