Gemini 3.7 Flash vs GPT-5.5
Gemini 3.7 Flash 
GPT-5.5 Verdict
- GPT-5.5 leads on quality: 66.6 vs 64.4.
- Gemini 3.7 Flash (medium) is stronger in coding, composite, knowledge, multimodal.
- GPT-5.5 is stronger in long context, reasoning, agents & tools, human preference, instruction following.
- Gemini 3.7 Flash (medium) is 7.5× cheaper ($1.50 vs $11.25 per 1M blended).
- Gemini 3.7 Flash (medium) streams 3.2× faster (282 vs 88 tokens per second).
| Metric | Gemini 3.7 Flash (medium) | GPT-5.5 |
|---|---|---|
| BenchLeader Index | 64.4 | 66.6 |
| Coding score | 68.3 | 57.3 |
| Composite score | 78.9 | 77.7 |
| Knowledge score | 75.2 | 73.7 |
| Long context score | 68.1 | 68.8 |
| Multimodal score | 69.4 | 64.9 |
| Reasoning score | 63.5 | 69.7 |
| Agents & tools score | – | 68.4 |
| Human preference score | – | 67.6 |
| Instruction following score | – | 73.3 |
| Blended price $/M | $1.50 | $11.25 |
| Output speed | 282 tok/s | 88 tok/s |
| Time to first answer | 5.3 s | 62.3 s |
| Context window | 1M | 1.1M |
| Terminal-Bench | – | 84.7% |
| SimpleBench | – | 69.0% |
| SciCode | 57.9% | – |
| Remote Labor Index | – | 6.3% |
| APEX-Agents | – | 38.5% |
| FrontierCode | – | 43.0% |
| Epoch Capabilities Index | – | 159.1 |
| LMArena Text | – | 1477 |
| LMArena Hard Prompts | – | 1498 |
| LMArena Coding | – | 1509 |
| LMArena WebDev | – | 1458 |
| LMArena Vision | – | 1296 |
| LMArena Agent | – | 3 |
| AA Intelligence Index | 39.6 | 38.6 |
| IFBench | – | 75.8% |
| AA-LCR | 83.0% | 84.3% |
| MMMU-Pro | 84.7% | 79.9% |
| AA-Omniscience | 23.7 | 20.5 |
| Terminal-Bench Hard | – | 60.6% |
| GPQA Diamond (AA) | 92.1% | 93.5% |
| Humanity's Last Exam (AA) | 39.0% | 45.8% |
| SciCode (AA) | 59.8% | 55.8% |
| τ²-Bench Telecom (AA) | – | 93.9% |
| HiL-Bench | – | 39.7% |
| EQ-Bench 4 | – | 1315 |
| Kagi LLM Benchmark | – | 88.8% |
| ARC-AGI-1 | 91.2% | – |
| ARC-AGI-2 | 63.8% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.