GPT-5.5 vs Kimi K3
GPT-5.5 
Kimi K3 Verdict
- GPT-5.5 and Kimi K3 are level on quality (66.6 vs 66.5).
- GPT-5.5 is stronger in agents & tools, composite, instruction following, knowledge, reasoning.
- Kimi K3 is stronger in coding, human preference, long context, multimodal, maths.
- Kimi K3 is 1.9× cheaper ($6.00 vs $11.25 per 1M blended).
- GPT-5.5 streams 2.4× faster (88 vs 37 tokens per second).
| Metric | GPT-5.5 | Kimi K3 |
|---|---|---|
| BenchLeader Index | 66.6 | 66.5 |
| Agents & tools score | 68.4 | 67.5 |
| Coding score | 57.3 | 64.9 |
| Composite score | 77.7 | 74.1 |
| Human preference score | 67.6 | 71.2 |
| Instruction following score | 73.3 | – |
| Knowledge score | 73.7 | 68.1 |
| Long context score | 68.8 | 71.1 |
| Multimodal score | 64.9 | 65.1 |
| Reasoning score | 69.7 | – |
| Maths score | – | 78.1 |
| Blended price $/M | $11.25 | $6.00 |
| Output speed | 88 tok/s | 37 tok/s |
| Time to first answer | 62.3 s | 57.7 s |
| Context window | 1.1M | 1.0M |
| Terminal-Bench | 84.7% | – |
| SimpleBench | 69.0% | – |
| SciCode | – | 58.7% |
| Remote Labor Index | 6.3% | – |
| APEX-Agents | 38.5% | 39.3% |
| FrontierCode | 43.0% | 44.2% |
| ProofBench | – | 87.0% |
| Epoch Capabilities Index | 159.1 | 157.6 |
| LMArena Text | 1477 | – |
| LMArena Hard Prompts | 1498 | – |
| LMArena Coding | 1509 | – |
| LMArena WebDev | 1458 | – |
| LMArena Vision | 1296 | – |
| LMArena Agent | 3 | – |
| LiveBench | – | 79.2% |
| LiveBench Reasoning | – | 90.7% |
| LiveBench Coding | – | 81.5% |
| LiveBench Agentic Coding | – | 62.2% |
| LiveBench Mathematics | – | 84.4% |
| LiveBench Data Analysis | – | 78.7% |
| LiveBench Language | – | 85.5% |
| AA Intelligence Index | 38.6 | 43.8 |
| IFBench | 75.8% | – |
| AA-LCR | 84.3% | 88.7% |
| MMMU-Pro | 79.9% | 80.5% |
| AA-Omniscience | 20.5 | 19.7 |
| Terminal-Bench Hard | 60.6% | – |
| GPQA Diamond (AA) | 93.5% | 93.5% |
| Humanity's Last Exam (AA) | 45.8% | 46.9% |
| SciCode (AA) | 55.8% | 59.5% |
| τ²-Bench Telecom (AA) | 93.9% | – |
| LegalBench | – | 86.0% |
| CorpFin | – | 71.6% |
| TaxEval | – | 75.7% |
| Terminal-Bench 2.1 (Vals) | – | 80.9% |
| SWE-bench (Vals) | – | 93.4% |
| GPQA Diamond (Vals) | – | 92.9% |
| Vals Index | – | 57.8 |
| HiL-Bench | 39.7% | – |
| EQ-Bench 4 | 1315 | 1339 |
| Kagi LLM Benchmark | 88.8% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.