Compare / Gemini 3.6 Flash vs GPT-5.4 Nano
Gemini 3.6 Flash vs GPT-5.4 Nano
27 results from 4 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 | GPT-5.4 Nano |
|---|---|---|---|
| BulletBenchmeasured by us | Cost per gameUS dollars | $0.0113 | $0.0050 |
| BulletBenchmeasured by us | Games lost on time% of games | 41.7% | 66.7% |
| BulletBenchmeasured by us | Invalid moves% of moves | 0.3% | 3.1% |
| BulletBenchmeasured by us | Ladder Eloladder Elo | 628 | 33 |
| BulletBenchmeasured by us | Median move timemilliseconds | 1.6 s | 1.1 s |
| BulletBenchmeasured by us | Cost per gameUS dollars | $0.0050 | $0.0025 |
| BulletBenchmeasured by us | Games lost on time% of games | 91.7% | 75.0% |
| BulletBenchmeasured by us | Invalid moves% of moves | 0.0% | 4.0% |
| BulletBenchmeasured by us | Ladder Eloladder Elo | 33 | 197 |
| BulletBenchmeasured by us | Median move timemilliseconds | 1.5 s | 1.0 s |
| Arena (formerly LMArena)reported | OverallArena rating | 1,470 | 1,418 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,472 | 1,405 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,469 | 1,335 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,500 | 1,438 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,473 | 1,387 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,482 | 1,401 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,472 | 1,362 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 69.2 | 56.1 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 34.0 | 20.7 |
| Artificial Analysisreported | GPQA Diamond% of questions | 92.8% | 81.7% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 40.8% | 28.3% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 80.0% | 76.7% |
| Artificial Analysisreported | SciCode% of problems | 53.4% | 47.2% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 77.5% | 60.7% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 7.1% | 0.5% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 29.9% | 27.4% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 66.2% | 11.7% |
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.