Gemini 3.8 Flash vs GPT-5.6 Sol
Gemini 3.8 Flash 
GPT-5.6 Sol Verdict
- GPT-5.6 Sol (high) leads on quality: 68.0 vs 63.5.
- Gemini 3.8 Flash (medium) is stronger in knowledge, long context, multimodal.
- GPT-5.6 Sol (high) is stronger in coding, composite, agents & tools, instruction following, reasoning.
- Gemini 3.8 Flash (medium) is 5.3× cheaper ($1.50 vs $8.00 per 1M blended).
- Gemini 3.8 Flash (medium) streams 4.6× faster (273 vs 59 tokens per second).
| Metric | Gemini 3.8 Flash (medium) | GPT-5.6 Sol (high) |
|---|---|---|
| BenchLeader Index | 63.5 | 68.0 |
| Coding score | 63.6 | 72.0 |
| Composite score | 79.4 | 82.5 |
| Knowledge score | 77.6 | 73.7 |
| Long context score | 68.7 | 67.4 |
| Multimodal score | 68.9 | 66.5 |
| Agents & tools score | – | 87.6 |
| Instruction following score | – | 67.6 |
| Reasoning score | – | 62.9 |
| Blended price $/M | $1.50 | $8.00 |
| Output speed | 273 tok/s | 59 tok/s |
| Time to first answer | 13.8 s | 31.4 s |
| Context window | 1M | 1M |
| SciCode | 54.4% | 56.9% |
| WeirdML | – | 88.8% |
| AA Intelligence Index | 40 | 42.5 |
| IFBench | – | 69.2% |
| AA-LCR | 84.0% | 81.7% |
| MMMU-Pro | 84.2% | 81.8% |
| AA-Omniscience | 28.6 | 20.4 |
| Terminal-Bench Hard | – | 62.1% |
| GPQA Diamond (AA) | 93.5% | 92.8% |
| Humanity's Last Exam (AA) | 42.1% | 46.0% |
| SciCode (AA) | 55.1% | 57.8% |
| τ²-Bench Telecom (AA) | – | 83.3% |
| EnigmaEval | – | 37.1% |
| ARC-AGI-1 | – | 97.0% |
| ARC-AGI-2 | – | 85.4% |
| ARC-AGI-3 | – | 2.2% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.