# DeepSeek V4 Pro vs GPT-5.1

> DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 48.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-pro-vs-gpt-5-1
- Last updated: 2026-10-10
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 7 categories and GPT-5.1 in 2 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.4 to 39.9.
- The biggest single-benchmark swing is ARC-AGI-2: 61.3% for DeepSeek V4 Pro and 17.6% for GPT-5.1.
- DeepSeek V4 Pro is cheaper at $0.43 / $0.87 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 400K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 54.2 | 48.8 |
| Rank | 28 | 47 |
| Context | 1M | 400K |
| Input $/M | $0.43 | $1.25 |
| Output $/M | $0.87 | $10 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Pro: 52.2 (#31)
- GPT-5.1: 46.9 (#59)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| SWE-bench Verified | 77.6% | 68% |
| LMArena WebDev | 1582 | 1395 |
| SciCode | 51% | 43.3% |
| WeirdML | 66.2% | 60.8% |
| LMArena Coding | 1470 | 1454 |
| ALE-Bench | 1,403 | 1,192 |
| FrontierCode | 28.6% | — |
| SWE-bench Verified (bash only) | — | 66% |
| GSO | — | 13.7% |
| LiveBench Coding | — | 72.5% |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.7 (#54)
- GPT-5.1: 32.5 (#55)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| Vending-Bench 2 | 3,285 | 1,473 |
| Terminal-Bench | — | 47.6% |
| APEX-Agents | 47.3% | — |
| DeepResearch Bench | — | 42.8% |
| LMArena Search | — | 1199 |

## Reasoning

- DeepSeek V4 Pro: 56.4 (#22)
- GPT-5.1: 39.9 (#50)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| ARC-AGI-2 | 61.3% | 17.6% |
| ARC-AGI-1 | 90.5% | 72.8% |
| CritPt | 18% | 4.9% |
| Chess Puzzles | 47% | 32% |
| LMArena Hard Prompts | 1461 | 1457 |
| Mystery Game Puzzles | 43% | 19% |
| DTBench | 93.9% | 90.1% |
| LMCA | 45.5% | 43.9% |
| Epoch Capabilities Index | 155.39 | 149.66 |
| ForecastBench | 56.1 | 58.1 |
| SimpleBench | — | 53.2% |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| EnigmaEval | — | 11.2% |
| LiveBench Reasoning | — | 95.8% |
| LiveBench Data Analysis | — | 72.1% |
| Surface Evolver Bench | 40% | — |
| LiveBench | — | 78.8% |

## Math

- DeepSeek V4 Pro: 64.6 (#24)
- GPT-5.1: 50.3 (#45)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 88.6% |
| LMArena Math | 1455 | 1446 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |
| Omni-MATH | — | 46.4% |
| LiveBench Math | — | 94.5% |
| FrontierMath (Feb 2025 set) | — | 31% |
| FrontierMath Tier 4 (v1) | — | 12.5% |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#28)
- GPT-5.1: 50.5 (#65)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 91.7% | 87.6% |
| SimpleQA Verified | 52.9% | 48% |
| Vectara Hallucination Rate | 8.6% | 10.9% |
| LMArena Expert | 1463 | 1470 |
| Humanity's Last Exam | — | 23.7% |
| MMLU-Pro | — | 57.9% |
| GPQA (HELM) | — | 44.2% |

## Multimodal

- DeepSeek V4 Pro: —
- GPT-5.1: 44.0 (#17)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| LMArena Vision | — | 1250 |
| VPCT | — | 58.7% |
| LMArena Document | — | 1403 |

## Multilingual

- DeepSeek V4 Pro: 54.5 (#44)
- GPT-5.1: 54.0 (#51)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1438 | 1431 |
| LMArena Chinese | 1486 | 1495 |
| LMArena French | 1473 | 1451 |
| LMArena German | 1458 | 1439 |
| LMArena Japanese | 1448 | 1453 |
| LMArena Korean | 1452 | 1401 |
| LMArena Russian | 1452 | 1435 |
| LMArena Spanish | 1457 | 1430 |

## Instruction Following

- DeepSeek V4 Pro: 76.3 (#44)
- GPT-5.1: 84.0 (#1)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1442 |
| LiveBench Instruction Following | — | 93.3% |
| IFEval | — | 93.5% |

## Long Context

- DeepSeek V4 Pro: 45.0 (#49)
- GPT-5.1: 47.6 (#14)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| CL-bench Life | 13.5% | 17.3% |
| LMArena Longer Query | 1459 | 1448 |
| CL-bench | — | 23.7% |

## Writing & Preference

- DeepSeek V4 Pro: 65.6 (#45)
- GPT-5.1: 64.7 (#53)

| Benchmark | DeepSeek V4 Pro | GPT-5.1 |
|---|---|---|
| LMArena Text | 1451 | 1443 |
| LMArena Creative Writing | 1446 | 1427 |
| LMArena Multi-Turn | 1467 | 1449 |
| EQ-Bench Creative Writing | 1553 | — |
| WildBench | — | 86.3% |
| EQ-Bench 4 | 1166 | — |
| LiveBench Language | — | 80.2% |

## FAQ

### Is DeepSeek V4 Pro better than GPT-5.1?

DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 48.8 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Pro or GPT-5.1?

DeepSeek V4 Pro is cheaper. It lists at $0.43 per million input tokens and $0.87 per million output tokens; GPT-5.1 lists at $1.25 and $10.

### Is DeepSeek V4 Pro or GPT-5.1 better for coding?

DeepSeek V4 Pro scores higher on coding benchmarks: 52.2 versus 46.9 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Pro does, with 1M tokens against 400K.

### How many benchmarks do DeepSeek V4 Pro and GPT-5.1 share?

37 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and GPT-5.1 has 63.