Compare / Gemini 3.6 Flash vs Kimi K3
Gemini 3.6 Flash vs Kimi K3
28 results from 6 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Gemini 3.6 Flash | Kimi K3 |
|---|---|---|---|
| BulletBenchmeasured by us | Cost per gameUS dollars | $0.0113 | $0.0540 |
| BulletBenchmeasured by us | Games lost on time% of games | 41.7% | 91.7% |
| BulletBenchmeasured by us | Invalid moves% of moves | 0.3% | 0.0% |
| BulletBenchmeasured by us | Ladder Eloladder Elo | 628 | 33 |
| BulletBenchmeasured by us | Median move timemilliseconds | 1.6 s | 2.6 s |
| APEX-Agentsreported | Tasks passed% of tasks | 46.9% | 50.6% |
| Arena (formerly LMArena)reported | Confirmed task successIPS effect estimate | -0.08 | 0.09 |
| Arena (formerly LMArena)reported | Praise over complaintIPS effect estimate | -0.09 | 0.07 |
| Arena (formerly LMArena)reported | SteerabilityIPS effect estimate | -0.04 | -0.01 |
| Arena (formerly LMArena)reported | Tool groundingIPS effect estimate | 0.00 | 0.00 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,470 | 1,472 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,472 | 1,489 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,469 | 1,458 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,500 | 1,533 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,473 | 1,486 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,482 | 1,488 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,472 | 1,475 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 69.2 | 76.2 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 34.0 | 43.6 |
| Artificial Analysisreported | GPQA Diamond% of questions | 92.8% | 93.5% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 40.8% | 46.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 80.0% | 88.7% |
| Artificial Analysisreported | SciCode% of problems | 53.4% | 59.5% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 77.5% | 85.0% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 7.1% | 12.6% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 29.9% | 46.0% |
| GDP.pdfreported | Professional document tasks% of rubric | 14.0% | 19.0% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 66.2% | 50.6% |
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.