Deepseek v4 Pro vs Gemini 3.5 Flash
Deepseek v4 Pro 
Gemini 3.5 Flash Verdict
- Gemini 3.5 Flash (medium) leads on quality: 62.2 vs 60.4.
- Deepseek v4 Pro (high) is stronger in coding, maths.
- Gemini 3.5 Flash (medium) is stronger in human preference, reasoning, agents & tools, composite, instruction following, knowledge, long context, multimodal.
| Metric | Deepseek v4 Pro (high) | Gemini 3.5 Flash (medium) |
|---|---|---|
| BenchLeader Index | 60.4 | 62.2 |
| Coding score | 59.3 | 58.7 |
| Human preference score | 66.0 | 67.5 |
| Maths score | 66.2 | – |
| Reasoning score | 65.9 | 67.0 |
| Agents & tools score | – | 63.3 |
| Composite score | – | 65.3 |
| Instruction following score | – | 69.4 |
| Knowledge score | – | 70.7 |
| Long context score | – | 60.7 |
| Multimodal score | – | 67.3 |
| Blended price $/M | – | $3.38 |
| Output speed | – | 234 tok/s |
| Time to first answer | – | 11.2 s |
| Context window | – | 1.0M |
| GPQA Diamond | 90.9% | – |
| OTIS Mock AIME | 95.6% | – |
| SciCode | 46.4% | – |
| WeirdML | 46.5% | – |
| LMArena Text | 1462 | 1474 |
| LMArena Hard Prompts | 1482 | 1493 |
| LMArena Coding | 1505 | 1502 |
| LMArena WebDev | 1581 | 1492 |
| LMArena Vision | – | 1306 |
| AA Intelligence Index | – | 33.6 |
| IFBench | – | 74.6% |
| AA-LCR | – | 74.3% |
| MMMU-Pro | – | 83.9% |
| AA-Omniscience | – | 20.8 |
| Terminal-Bench Hard | – | 39.4% |
| GPQA Diamond (AA) | – | 92.1% |
| Humanity's Last Exam (AA) | – | 41.3% |
| τ²-Bench Telecom (AA) | – | 95.6% |
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 3.5 Flash: questions
- Is Deepseek v4 Pro better than Gemini 3.5 Flash?
- Gemini 3.5 Flash (medium) leads on quality: 62.2 vs 60.4. The BenchLeader Index combines every independent quality benchmark; Gemini 3.5 Flash (medium) is ahead overall as of 2026-09-13, but check the category scores for your use.
- Is Deepseek v4 Pro better than Gemini 3.5 Flash for coding?
- Deepseek v4 Pro scores higher in coding (59 vs 59 on the category index, where 50 is average).