Compare / Claude Sonnet 4.6 vs Gemini 3.6 Flash
Claude Sonnet 4.6 vs Gemini 3.6 Flash
23 results from 6 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Sonnet 4.6 | Gemini 3.6 Flash |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 43.0% | 46.9% |
| Arena (formerly LMArena)reported | OverallArena rating | 1,478 | 1,453 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,461 | 1,470 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,482 | 1,472 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,449 | 1,469 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,506 | 1,500 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,475 | 1,473 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,472 | 1,482 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,455 | 1,472 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 63.0 | 69.2 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 30.1 | 34.0 |
| Artificial Analysisreported | GPQA Diamond% of questions | 87.5% | 92.8% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 33.6% | 40.8% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 80.0% | 80.0% |
| Artificial Analysisreported | SciCode% of problems | 50.1% | 53.4% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 71.2% | 77.5% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 3.0% | 7.1% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 34.4% | 29.9% |
| GDP.pdfreported | Professional document tasks% of rubric | 18.0% | 14.0% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 35.5% | 66.2% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 2.0% | 9.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.41 | 0.33 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 64.5 | 69.8 |
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.