Deepseek v4 Pro vs Gemini 3.8 Flash
Deepseek v4 Pro 
Gemini 3.8 Flash Verdict
- Deepseek v4 Pro (high) and Gemini 3.8 Flash are level on quality (60.4 vs 61.1).
- Deepseek v4 Pro (high) is stronger in coding, human preference, maths, reasoning.
- Gemini 3.8 Flash is stronger in agents & tools, composite, knowledge, long context, multimodal.
| Metric | Deepseek v4 Pro (high) | Gemini 3.8 Flash |
|---|---|---|
| BenchLeader Index | 60.4 | 61.1 |
| Coding score | 59.3 | – |
| Human preference score | 66.0 | – |
| Maths score | 66.2 | 57.0 |
| Reasoning score | 65.9 | – |
| Agents & tools score | – | 57.2 |
| Composite score | – | 74.1 |
| Knowledge score | – | 68.1 |
| Long context score | – | 64.9 |
| Multimodal score | – | 69.9 |
| Blended price $/M | – | $1.50 |
| Output speed | – | 277 tok/s |
| Time to first answer | – | 12.8 s |
| Context window | – | 1.0M |
| GPQA Diamond | 90.9% | – |
| OTIS Mock AIME | 95.6% | – |
| SciCode | 46.4% | – |
| WeirdML | 46.5% | – |
| ProofBench | – | 48.0% |
| LMArena Text | 1462 | – |
| LMArena Hard Prompts | 1482 | – |
| LMArena Coding | 1505 | – |
| LMArena WebDev | 1581 | – |
| AA Intelligence Index | – | 41.2 |
| AA-LCR | – | 81.3% |
| MMMU-Pro | – | 85.6% |
| AA-Omniscience | – | 29.6 |
| GPQA Diamond (AA) | – | 95.3% |
| Humanity's Last Exam (AA) | – | 47.8% |
| SciCode (AA) | – | 56.6% |
| PRBench Finance | – | 49.5% |
| PRBench Legal | – | 48.9% |
| HiL-Bench | – | 41.5% |
Data as of 2026-09-13. Best configuration of each model; every score links to its source on the model pages.
Deepseek v4 Pro vs Gemini 3.8 Flash: questions
- Is Deepseek v4 Pro better than Gemini 3.8 Flash?
- Deepseek v4 Pro (high) and Gemini 3.8 Flash are level on quality (60.4 vs 61.1). The BenchLeader Index combines every independent quality benchmark; Gemini 3.8 Flash is ahead overall as of 2026-09-13, but check the category scores for your use.