Compare / Gemini 3.8 Flash vs GPT-5.5
Gemini 3.8 Flash vs GPT-5.5
27 results from 7 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Gemini 3.8 Flash | GPT-5.5 |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 64.3% | 55.1% |
| Arena (formerly LMArena)reported | Confirmed task successIPS effect estimate | 0.09 | -0.03 |
| Arena (formerly LMArena)reported | Praise over complaintIPS effect estimate | 0.05 | 0.08 |
| Arena (formerly LMArena)reported | SteerabilityIPS effect estimate | 0.04 | 0.07 |
| Arena (formerly LMArena)reported | Tool groundingIPS effect estimate | 0.00 | 0.00 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,477 | 1,486 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,482 | 1,489 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,482 | 1,452 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,529 | 1,513 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,488 | 1,478 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,492 | 1,481 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,486 | 1,470 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 76.3 | 74.9 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 40.9 | 38.4 |
| Artificial Analysisreported | GPQA Diamond% of questions | 95.3% | 93.5% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 47.8% | 45.8% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 84.0% | 84.3% |
| Artificial Analysisreported | SciCode% of problems | 56.6% | 56.1% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 87.6% | 84.3% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 19.7% | 14.6% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 45.8% | 39.0% |
| GDP.pdfreported | Professional document tasks% of rubric | 23.4% | 26.0% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 69.7% | 63.0% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 13.0% | 13.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.34 | 0.41 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 78.5 | 69.7 |
| Vending-Bench 2reported | Money after a yearUS dollars | $5,093.79 | $7,523.84 |
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.