Claude Opus 4.8 vs Gemini 3.8 Flash
Claude Opus 4.8 
Gemini 3.8 Flash Verdict
- Claude Opus 4.8 leads on quality: 64.7 vs 63.5.
- Claude Opus 4.8 is stronger in agents & tools, coding, composite, human preference, instruction following, reasoning.
- Gemini 3.8 Flash (medium) is stronger in knowledge, long context, multimodal.
- Gemini 3.8 Flash (medium) is 6.7× cheaper ($1.50 vs $10.00 per 1M blended).
- Gemini 3.8 Flash (medium) streams 4.7× faster (273 vs 58 tokens per second).
| Metric | Claude Opus 4.8 | Gemini 3.8 Flash (medium) |
|---|---|---|
| BenchLeader Index | 64.7 | 63.5 |
| Agents & tools score | 65.0 | – |
| Coding score | 65.7 | 63.6 |
| Composite score | 81.9 | 79.4 |
| Human preference score | 65.6 | – |
| Instruction following score | 61.6 | – |
| Knowledge score | 66.3 | 77.6 |
| Long context score | 65.3 | 68.7 |
| Multimodal score | 64.4 | 68.9 |
| Reasoning score | 64.2 | – |
| Blended price $/M | $10.00 | $1.50 |
| Output speed | 58 tok/s | 273 tok/s |
| Time to first answer | 29.5 s | 13.8 s |
| Context window | 1M | 1M |
| SimpleBench | 64.8% | – |
| SciCode | – | 54.4% |
| Remote Labor Index | 8.3% | – |
| FrontierCode | 46.5% | – |
| GSO-Bench | 47.1% | – |
| Epoch Capabilities Index | 158.3 | – |
| LMArena Text | 1473 | – |
| LMArena Hard Prompts | 1503 | – |
| LMArena Coding | 1527 | – |
| LMArena WebDev | 1540 | – |
| LMArena Vision | 1289 | – |
| AA Intelligence Index | 42.0 | 40 |
| IFBench | 62.2% | – |
| AA-LCR | 77.7% | 84.0% |
| MMMU-Pro | – | 84.2% |
| AA-Omniscience | 28.8 | 28.6 |
| Terminal-Bench Hard | 58.3% | – |
| GPQA Diamond (AA) | 92.0% | 93.5% |
| Humanity's Last Exam (AA) | 48.7% | 42.1% |
| SciCode (AA) | 54.4% | 55.1% |
| τ²-Bench Telecom (AA) | 94.4% | – |
| LiveCodeBench | 87.8% | – |
| MMLU-Pro | 89.6% | – |
| LegalBench | 83.6% | – |
| CorpFin | 66.7% | – |
| TaxEval | 75.6% | – |
| Terminal-Bench 2.1 (Vals) | 71.9% | – |
| SWE-bench (Vals) | 88.6% | – |
| GPQA Diamond (Vals) | 92.4% | – |
| Vals Index | 60.9 | – |
| HiL-Bench | 35.3% | – |
| EQ-Bench 4 | 1281 | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.