Claude Sonnet 5.5 vs Kimi K3
Claude Sonnet 5.5 
Kimi K3 Verdict
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 65.5.
- Claude Sonnet 5.5 (max) is stronger in agents & tools, coding, knowledge, maths, reasoning.
- Kimi K3 (max) is stronger in composite, long context, human preference, multimodal.
- Claude Sonnet 5.5 (max) is 1.5× cheaper ($4.00 vs $6.00 per 1M blended).
- Claude Sonnet 5.5 (max) streams 3.4× faster (141 vs 41 tokens per second).
| Metric | Claude Sonnet 5.5 (max) | Kimi K3 (max) |
|---|---|---|
| BenchLeader Index | 67.6 | 65.5 |
| Agents & tools score | 69.5 | 66.2 |
| Coding score | 69.8 | 64.7 |
| Composite score | 72.6 | 79.0 |
| Knowledge score | 65.4 | 63.2 |
| Long context score | 66.6 | 69.7 |
| Maths score | 72.9 | 63.8 |
| Reasoning score | 81.1 | 66.6 |
| Human preference score | – | 68.6 |
| Multimodal score | – | 63.7 |
| Blended price $/M | $4.00 | $6.00 |
| Output speed | 141 tok/s | 41 tok/s |
| Time to first answer | 487.1 s | 52.2 s |
| Context window | 1M | 1.0M |
| GPQA Diamond | 95.6% | 93.1% |
| FrontierMath Tiers 1–3 | 88.8% | 72.2% |
| FrontierMath Tier 4 | 80.5% | 39.0% |
| OTIS Mock AIME | 100.0% | 97.2% |
| SimpleQA Verified | 46.5% | 50.6% |
| Terminal-Bench | 61.8% | – |
| SimpleBench | – | 60.7% |
| SciCode | 61.0% | 59.5% |
| WeirdML | – | 82.6% |
| APEX-Agents | 75.5% | – |
| ProofBench | 100.0% | – |
| LMArena Text | – | 1488 |
| LMArena Hard Prompts | – | 1517 |
| LMArena Coding | – | 1540 |
| LMArena WebDev | – | 1654 |
| LMArena Agent | 12 | 3.8 |
| LiveBench | 75.7% | – |
| LiveBench Reasoning | 91.6% | – |
| LiveBench Coding | 91.4% | – |
| LiveBench Agentic Coding | 56.3% | – |
| LiveBench Mathematics | 96.1% | – |
| LiveBench Data Analysis | 59.5% | – |
| LiveBench Language | 78.0% | – |
| LiveBench Instruction Following | 56.8% | – |
| AA Intelligence Index v4.3.2 | 56 | 43.6 |
| AA-LCR | 82.7% | 88.7% |
| MMMU-Pro | – | 80.5% |
| AA-Omniscience | 32.3 | 19.7 |
| GPQA Diamond (AA) | – | 93.5% |
| Humanity's Last Exam (AA) | 55.0% | 46.9% |
| SciCode (AA) | 61.0% | 59.5% |
| LiveCodeBench | – | 87.2% |
| MMLU-Pro | – | 88.0% |
| IOI | – | 48.9% |
| MCP Atlas | – | 82.3% |
| ARC-AGI-1 | – | 94.5% |
| ARC-AGI-2 | – | 60.4% |
| CritPt | 31.4% | 23.4% |
| GDPval-AA v2.1 | 67.0% | 51.7% |
| τ³-Banking (AA) | – | 46.0% |
| ITBench SRE (AA) | – | 47.7% |
| Analyst Agent (AA) | 57.5% | 38.8% |
| APEX-Agents (AA) | – | 41.3% |
| BioMysteryBench | – | 71.5% |
| Code Migration | – | 16.1% |
| CyberBench | – | 75.2% |
| Excel Modeling Benchmark | – | 66.4% |
| Harvey's Legal Agent Benchmark | – | 10.8% |
| Public Benefits Bench | – | 68.3% |
| Tax Agent Bench | – | 68.7% |
| Terminal-Bench 4.0 (Vals) | – | 19.7% |
| Terminal-Bench Science | – | 2.9% |
| Time Horizon Index: KSP | – | 10.5% |
| Vibe Code Bench 1-100 | – | 18.2% |
| FORTRESS | – | 26.6% |
| LMArena Maths | – | 1498 |
| LMArena Creative Writing | – | 1460 |
| LMArena Instruction Following | – | 1488 |
| LMArena Multi-turn | – | 1500 |
| LMArena Longer Queries | – | 1504 |
| Chess Puzzles | – | 39.0% |
| Mystery Game Puzzles | 65.0% | 26.0% |
| Surface Evolver Bench | – | 93.0% |
| DeepSWE v1.1 | – | 68.5% |
| LMCA | – | 52.7% |
| DTBench | – | 91.2% |
| ForecastBench | – | 61.1% |
| CursorBench | 55.5% | – |
| ALE-Bench | – | 1524.5 |
| GDP.pdf | – | 19.0% |
| FrontierSWE | 61.9% | 25.9% |
| Terminal-Bench 4.0 (AA) | 63.6% | 12.6% |
| Terminal-Bench 2.1 (AA) | – | 85.0% |
| AutomationBench | 71.8% | 58.3% |
| GDP.pdf | 25.8% | 22.0% |
| MLCR | 75.0% | 38.3% |
| Harvey LAB | 2.8% | 5.3% |
| EnterpriseOps-Gym | – | 45.3% |
| AA-Omniscience: accuracy | 54.0% | 47.6% |
| AA-Omniscience: non-hallucination | 53.0% | 46.8% |
| AA-Briefcase v1.1 | 1823 | 1501 |
| 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.
Claude Sonnet 5.5 vs Kimi K3: questions
- Is Claude Sonnet 5.5 better than Kimi K3?
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 65.5. The BenchLeader Index combines every independent quality benchmark; Claude Sonnet 5.5 (max) is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is Claude Sonnet 5.5 better than Kimi K3 for coding?
- Claude Sonnet 5.5 scores higher in coding (70 vs 65 on the category index, where 50 is average).
- Is Claude Sonnet 5.5 better than Kimi K3 for agentic tasks?
- Claude Sonnet 5.5 scores higher in agentic tasks (70 vs 66 on the category index, where 50 is average).
- Which is cheaper, Claude Sonnet 5.5 or Kimi K3?
- Claude Sonnet 5.5 is cheaper: $4.00 against $6.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Sonnet 5.5 or Kimi K3?
- Claude Sonnet 5.5 streams faster: 141 against 41 output tokens per second.
- Which has the larger context window?
- Kimi K3 accepts more context: 1.0M against 1M tokens.