Compare / Kimi K2.6 vs Kimi K3
Kimi K2.6 vs Kimi K3
19 results from 5 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Kimi K2.6 | Kimi K3 |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,456 | 1,472 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,455 | 1,489 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,433 | 1,458 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,503 | 1,533 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,456 | 1,486 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,461 | 1,488 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,441 | 1,475 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 61.8 | 76.2 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 27.0 | 43.6 |
| Artificial Analysisreported | GPQA Diamond% of questions | 91.1% | 93.5% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 37.5% | 46.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 81.0% | 88.7% |
| Artificial Analysisreported | SciCode% of problems | 51.5% | 59.5% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 65.9% | 85.0% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 0.5% | 12.6% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 23.3% | 46.0% |
| GDP.pdfreported | Professional document tasks% of rubric | 12.0% | 19.0% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 34.9% | 50.6% |
| Vending-Bench 2reported | Money after a yearUS dollars | $6,204.57 | $5,165.03 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.