Claude Opus 4.6 vs Deepseek v4 Pro
Claude Opus 4.6 
Deepseek v4 Pro Verdict
- Claude Opus 4.6 (thinking) leads on quality: 62.5 vs 60.4.
- Claude Opus 4.6 (thinking) is stronger in coding, knowledge, reasoning.
- Deepseek v4 Pro (high) is stronger in maths, human preference.
| Metric | Claude Opus 4.6 (thinking) | Deepseek v4 Pro (high) |
|---|---|---|
| BenchLeader Index | 62.5 | 60.4 |
| Coding score | 63.6 | 59.3 |
| Knowledge score | 62.2 | – |
| Maths score | 65.1 | 66.2 |
| Reasoning score | 75.6 | 65.9 |
| Human preference score | – | 66.0 |
| Blended price $/M | $10.00 | – |
| Output speed | 38 tok/s | – |
| Time to first answer | 2.0 s | – |
| Context window | 1M | – |
| GPQA Diamond | – | 90.9% |
| OTIS Mock AIME | – | 95.6% |
| SciCode | – | 46.4% |
| WeirdML | – | 46.5% |
| LMArena Text | – | 1462 |
| LMArena Hard Prompts | – | 1482 |
| LMArena Coding | – | 1505 |
| LMArena WebDev | – | 1581 |
| AIME (Vals) | 95.6% | – |
| LiveCodeBench | 84.7% | – |
| MMLU-Pro | 89.1% | – |
| LegalBench | 85.3% | – |
| CorpFin | 67.0% | – |
| TaxEval | 76.0% | – |
| MedQA | 95.4% | – |
| SWE-bench (Vals) | 78.2% | – |
| GPQA Diamond (Vals) | 89.7% | – |
| SWE-Bench Pro | 51.9% | – |
| Kagi LLM Benchmark | 83.6% | – |
Data as of 2026-09-13. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 4.6 vs Deepseek v4 Pro: questions
- Is Claude Opus 4.6 better than Deepseek v4 Pro?
- Claude Opus 4.6 (thinking) leads on quality: 62.5 vs 60.4. The BenchLeader Index combines every independent quality benchmark; Claude Opus 4.6 (thinking) is ahead overall as of 2026-09-13, but check the category scores for your use.
- Is Claude Opus 4.6 better than Deepseek v4 Pro for coding?
- Claude Opus 4.6 scores higher in coding (64 vs 59 on the category index, where 50 is average).