Deepseek v4 Pro vs GPT-5.6 Terra
Deepseek v4 Pro 
GPT-5.6 Terra Verdict
- GPT-5.6 Terra (xhigh) leads on quality: 62.6 vs 60.4.
- Deepseek v4 Pro (high) is stronger in reasoning.
- GPT-5.6 Terra (xhigh) is stronger in coding, human preference, maths, agents & tools, composite, instruction following, knowledge, long context, multimodal.
| Metric | Deepseek v4 Pro (high) | GPT-5.6 Terra (xhigh) |
|---|---|---|
| BenchLeader Index | 60.4 | 62.6 |
| Coding score | 59.3 | 61.5 |
| Human preference score | 66.0 | 66.5 |
| Maths score | 66.2 | 71.0 |
| Reasoning score | 65.9 | 58.6 |
| Agents & tools score | – | 66.9 |
| Composite score | – | 70.6 |
| Instruction following score | – | 62.1 |
| Knowledge score | – | 59.9 |
| Long context score | – | 63.5 |
| Multimodal score | – | 62.4 |
| Blended price $/M | – | $4.50 |
| Output speed | – | 97 tok/s |
| Time to first answer | – | 24.9 s |
| Context window | – | 1.1M |
| GPQA Diamond | 90.9% | – |
| OTIS Mock AIME | 95.6% | – |
| SimpleBench | – | 48.9% |
| SciCode | 46.4% | 51.6% |
| WeirdML | 46.5% | – |
| ProofBench | – | 74.0% |
| LMArena Text | 1462 | 1466 |
| LMArena Hard Prompts | 1482 | 1489 |
| LMArena Coding | 1505 | 1518 |
| LMArena WebDev | 1581 | 1521 |
| LMArena Vision | – | 1269 |
| LMArena Agent | – | 1.5 |
| AA Intelligence Index | – | 38.2 |
| IFBench | – | 66.3% |
| AA-LCR | – | 79.0% |
| MMMU-Pro | – | 79.5% |
| AA-Omniscience | – | -3.0 |
| Terminal-Bench Hard | – | 62.9% |
| GPQA Diamond (AA) | – | 90.8% |
| Humanity's Last Exam (AA) | – | 41.9% |
| SciCode (AA) | – | 52.3% |
| τ²-Bench Telecom (AA) | – | 80.4% |
| LiveCodeBench | – | 85.9% |
| MMLU-Pro | – | 86.7% |
| CorpFin | – | 65.3% |
| TaxEval | – | 76.2% |
| GPQA Diamond (Vals) | – | 90.9% |
| ARC-AGI-1 | – | 94.0% |
| ARC-AGI-2 | – | 74.2% |
| ARC-AGI-3 | – | 0.7% |
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-5.6 Terra: questions
- Is Deepseek v4 Pro better than GPT-5.6 Terra?
- GPT-5.6 Terra (xhigh) leads on quality: 62.6 vs 60.4. The BenchLeader Index combines every independent quality benchmark; GPT-5.6 Terra (xhigh) is ahead overall as of 2026-09-13, but check the category scores for your use.
- Is Deepseek v4 Pro better than GPT-5.6 Terra for coding?
- GPT-5.6 Terra scores higher in coding (62 vs 59 on the category index, where 50 is average).