Gemini 3.8 Flash vs GPT-5.4
Gemini 3.8 Flash 
GPT-5.4 Verdict
- Gemini 3.8 Flash (medium) and GPT-5.4 are level on quality (63.5 vs 63.4).
- Gemini 3.8 Flash (medium) is stronger in coding, composite, knowledge, long context, multimodal.
- GPT-5.4 is stronger in agents & tools, human preference, instruction following, reasoning.
- Gemini 3.8 Flash (medium) is 3.8× cheaper ($1.50 vs $5.63 per 1M blended).
- Gemini 3.8 Flash (medium) streams 2.1× faster (273 vs 130 tokens per second).
| Metric | Gemini 3.8 Flash (medium) | GPT-5.4 |
|---|---|---|
| BenchLeader Index | 63.5 | 63.4 |
| Coding score | 63.6 | 51.6 |
| Composite score | 79.4 | 78.1 |
| Knowledge score | 77.6 | 66.8 |
| Long context score | 68.7 | 67.6 |
| Multimodal score | 68.9 | 64.0 |
| Agents & tools score | – | 66.1 |
| Human preference score | – | 64.5 |
| Instruction following score | – | 71.7 |
| Reasoning score | – | 62.3 |
| Blended price $/M | $1.50 | $5.63 |
| Output speed | 273 tok/s | 130 tok/s |
| Time to first answer | 13.8 s | 126.4 s |
| Context window | 1M | 1.1M |
| Terminal-Bench | – | 81.8% |
| SciCode | 54.4% | – |
| APEX-Agents | – | 34.9% |
| Epoch Capabilities Index | – | 156.9 |
| LMArena Text | – | 1466 |
| LMArena Hard Prompts | – | 1488 |
| LMArena Coding | – | 1514 |
| LMArena WebDev | – | 1386 |
| LMArena Vision | – | 1293 |
| AA Intelligence Index | 40 | 39.0 |
| IFBench | – | 74.0% |
| AA-LCR | 84.0% | 82.0% |
| MMMU-Pro | 84.2% | 78.4% |
| AA-Omniscience | 28.6 | 5.8 |
| Terminal-Bench Hard | – | 57.6% |
| GPQA Diamond (AA) | 93.5% | 92.0% |
| Humanity's Last Exam (AA) | 42.1% | 43.7% |
| SciCode (AA) | 55.1% | – |
| τ²-Bench Telecom (AA) | – | 87.1% |
| HiL-Bench | – | 9.7% |
| EQ-Bench 4 | – | 1272 |
| Kagi LLM Benchmark | – | 63.8% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.