Deepseek v4 Pro vs GPT-5
Deepseek v4 Pro 
GPT-5 Verdict
- Deepseek v4 Pro (high) and GPT-5 are level on quality (60.4 vs 59.7).
- Deepseek v4 Pro (high) is stronger in human preference, reasoning.
- GPT-5 is stronger in coding, maths, agents & tools, composite, instruction following, knowledge, long context, multimodal.
| Metric | Deepseek v4 Pro (high) | GPT-5 |
|---|---|---|
| BenchLeader Index | 60.4 | 59.7 |
| Coding score | 59.3 | 60.9 |
| Human preference score | 66.0 | 61.7 |
| Maths score | 66.2 | 72.8 |
| Reasoning score | 65.9 | 64.2 |
| Agents & tools score | – | 50.0 |
| Composite score | – | 52.8 |
| Instruction following score | – | 64.0 |
| Knowledge score | – | 60.8 |
| Long context score | – | 63.0 |
| Multimodal score | – | 58.9 |
| Blended price $/M | – | $3.44 |
| Output speed | – | 106 tok/s |
| Time to first answer | – | 54.4 s |
| Context window | – | 400k |
| GPQA Diamond | 90.9% | – |
| OTIS Mock AIME | 95.6% | – |
| Terminal-Bench | – | 49.6% |
| SciCode | 46.4% | 42.9% |
| Remote Labor Index | – | 1.7% |
| WeirdML | 46.5% | 39.8% |
| APEX-Agents | – | 18.3% |
| Epoch Capabilities Index | – | 150 |
| LMArena Text | 1462 | 1427 |
| LMArena Hard Prompts | 1482 | 1449 |
| LMArena Coding | 1505 | 1462 |
| LMArena WebDev | 1581 | – |
| LMArena Vision | – | 1232 |
| AA Intelligence Index | – | 23.0 |
| IFBench | – | 73.1% |
| AA-LCR | – | 78.2% |
| MMMU-Pro | – | 74.2% |
| AA-Omniscience | – | -8.7 |
| Terminal-Bench Hard | – | 32.6% |
| GPQA Diamond (AA) | – | 85.3% |
| Humanity's Last Exam (AA) | – | 28.5% |
| τ²-Bench Telecom (AA) | – | 84.8% |
| PRBench Finance | – | 51.3% |
| PRBench Legal | – | 49.0% |
| VISTA | – | 49.7% |
| MultiNRC | – | 52.1% |
| TutorBench | – | 55.3% |
| Kagi LLM Benchmark | – | 72.7% |
| IFEval (HELM) | – | 87.5% |
| Omni-MATH (HELM) | – | 64.7% |
| WildBench (HELM) | – | 85.7% |
| MMLU-Pro (HELM) | – | 86.3% |
| GPQA Diamond (HELM) | – | 79.1% |
| HELM Capabilities mean | – | 80.7% |
| Aider Polyglot | – | 88.0% |
| SWE-bench Verified (any scaffold) | – | 75.6% |
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 GPT-5: questions
- Is Deepseek v4 Pro better than GPT-5?
- Deepseek v4 Pro (high) and GPT-5 are level on quality (60.4 vs 59.7). The BenchLeader Index combines every independent quality benchmark; Deepseek v4 Pro (high) is ahead overall as of 2026-09-13, but check the category scores for your use.
- Is Deepseek v4 Pro better than GPT-5 for coding?
- GPT-5 scores higher in coding (61 vs 59 on the category index, where 50 is average).