Model comparison
Claude Opus 4.1 vs DeepSeek V4 Pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 41.0 on the Noometry Index.
Last verified . 32 shared benchmarks.
DeepSeekSummary
- They share 32 benchmarks with published results for both. Claude Opus 4.1 scores higher in 1 category and DeepSeek V4 Pro in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.8 to 22.3.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 12.6% for Claude Opus 4.1 and 64.6% for DeepSeek V4 Pro.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $15 / $75 for Claude Opus 4.1.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 200K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4.1 | DeepSeek V4 Pro | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 41.0 | 54.3 |
| Released | 2025-08-05 | 2026-04-24 |
| Weights | Proprietary | Open |
| Context window | 200K | 1M |
| Max output | 32K | 393K |
| Input $ / M tokens | $15 | $0.66 |
| Output $ / M tokens | $75 | $1.98 |
| Results tracked | 48 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
Claude Opus 4.1: 44.4 (#73), DeepSeek V4 Pro: 52.4 (#34)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| SWE-bench Verified | 73.3% | 77.6% |
| LMArena WebDev | 1390 | 1582 |
| WeirdML | 45.9% | 66.2% |
| LMArena Coding | 1479 | 1470 |
| ALE-Bench | 674.77 | 1,403 |
| FrontierCode | — | 28.6% |
| SciCode | — | 51% |
| AlgoTune | 1.34 | — |
Agentic & Tool Use Claude Opus 4.1 leads
Claude Opus 4.1: 35.0 (#41), DeepSeek V4 Pro: 32.8 (#58)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| Terminal-Bench | 38% | — |
| APEX-Agents | — | 47.3% |
| GDPval | 43.6% | — |
| Cybench | 42% | — |
| DeepResearch Bench | 48.3% | — |
| LMArena Search | 1148 | — |
| METR Time Horizons | 66.8% | — |
| Vending-Bench 2 | — | 3,285 |
Reasoning DeepSeek V4 Pro leads
Claude Opus 4.1: 32.2 (#76), DeepSeek V4 Pro: 56.5 (#24)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| Chess Puzzles | 7% | 47% |
| LMArena Hard Prompts | 1443 | 1461 |
| Mystery Game Puzzles | 21% | 43% |
| DTBench | 80% | 93.9% |
| LMCA | 37.1% | 45.5% |
| Epoch Capabilities Index | 144.12 | 155.31 |
| ForecastBench | 62 | 56.1 |
| ARC-AGI-2 | — | 61.3% |
| SimpleBench | 60% | — |
| Kagi LLM Benchmark | — | 53.5% |
| NYT Connections (extended) | — | 91.3% |
| ARC-AGI-1 | — | 90.5% |
| CritPt | — | 18% |
| EnigmaEval | 7.2% | — |
| EBR-Bench | 7.9% | — |
| Surface Evolver Bench | — | 40% |
Math DeepSeek V4 Pro leads
Claude Opus 4.1: 22.3 (#277), DeepSeek V4 Pro: 64.8 (#30)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | 12.6% | 64.6% |
| FrontierMath Tier 4 | 2.4% | 26.8% |
| OTIS Mock AIME 2024-2025 | 68.9% | 98.6% |
| LMArena Math | 1431 | 1455 |
| MathArena Final-Answer Competitions | — | 76.6% |
| ProofBench | — | 50% |
| FrontierMath (Feb 2025 set) | 7.2% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge DeepSeek V4 Pro leads
Claude Opus 4.1: 42.0 (#101), DeepSeek V4 Pro: 59.5 (#31)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| GPQA Diamond | 77.3% | 91.7% |
| Vectara Hallucination Rate | 11.8% | 8.6% |
| LMArena Expert | 1439 | 1464 |
| Humanity's Last Exam | 11.5% | — |
| SimpleQA Verified | — | 52.9% |
| Confabulations | 17.1% | — |
Multimodal Not comparable
Claude Opus 4.1: 26.8 (#119), DeepSeek V4 Pro: —
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| VPCT | 35% | — |
Multilingual DeepSeek V4 Pro leads
Claude Opus 4.1: 52.0 (#95), DeepSeek V4 Pro: 54.4 (#45)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | 1405 | 1439 |
| LMArena Chinese | 1427 | 1486 |
| LMArena French | 1431 | 1472 |
| LMArena German | 1413 | 1458 |
| LMArena Japanese | 1378 | 1445 |
| LMArena Korean | 1380 | 1447 |
| LMArena Russian | 1422 | 1453 |
| LMArena Spanish | 1448 | 1458 |
Instruction Following Too close to call
Claude Opus 4.1: 75.6 (#58), DeepSeek V4 Pro: 76.1 (#47)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | 1435 | 1448 |
Long Context Too close to call
Claude Opus 4.1: 44.5 (#63), DeepSeek V4 Pro: 45.0 (#51)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| LMArena Longer Query | 1455 | 1458 |
| CL-bench Life | — | 13.5% |
Writing & Preference DeepSeek V4 Pro leads
Claude Opus 4.1: 62.4 (#74), DeepSeek V4 Pro: 65.5 (#46)
| Benchmark | Claude Opus 4.1 | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | 1419 | 1451 |
| LMArena Creative Writing | 1412 | 1446 |
| LMArena Multi-Turn | 1444 | 1467 |
| Short-Story Creative Writing | 84.7% | — |
| EQ-Bench Creative Writing | — | 1553 |
| EQ-Bench 4 | — | 1166 |
Frequently asked questions
Is Claude Opus 4.1 better than DeepSeek V4 Pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 41.0 on the Noometry Index.
Which is cheaper, Claude Opus 4.1 or DeepSeek V4 Pro?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Claude Opus 4.1 lists at $15 and $75.
Is Claude Opus 4.1 or DeepSeek V4 Pro better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 44.4 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 200K.
How many benchmarks do Claude Opus 4.1 and DeepSeek V4 Pro share?
32 benchmarks have published results for both models. Claude Opus 4.1 has 48 scored results on Noometry and DeepSeek V4 Pro has 48.