# DeepSeek V4 Pro vs GPT-4

> DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-pro-vs-gpt-4
- Last updated: 2026-10-10
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and GPT-4 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.8 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.6% for DeepSeek V4 Pro and 1.1% for GPT-4.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $30 / $60 for GPT-4.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 8K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 54.3 | 29.1 |
| Rank | 31 | 316 |
| Context | 1M | 8K |
| Input $/M | $0.66 | $30 |
| Output $/M | $1.98 | $60 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Pro: 52.4 (#34)
- GPT-4: 31.6 (#283)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| WeirdML | 66.2% | 12.4% |
| LMArena Coding | 1470 | 1254 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| BigCodeBench Instruct | — | 46% |
| BigCodeBench Complete | — | 57.2% |
| ALE-Bench | 1,403 | — |
| HumanEval+ | — | 79.3% |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.8 (#58)
- GPT-4: —

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| APEX-Agents | 47.3% | — |
| METR Time Horizons | — | 36.1% |
| Vending-Bench 2 | 3,285 | — |

## Reasoning

- DeepSeek V4 Pro: 56.5 (#24)
- GPT-4: 17.8 (#289)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| Chess Puzzles | 47% | 4% |
| LMArena Hard Prompts | 1461 | 1241 |
| Mystery Game Puzzles | 43% | 12% |
| DTBench | 93.9% | 62.7% |
| LMCA | 45.5% | 17.1% |
| Epoch Capabilities Index | 155.31 | 125.89 |
| ForecastBench | 56.1 | 57.8 |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| CritPt | 18% | — |
| Surface Evolver Bench | 40% | — |
| BIG-Bench Hard | — | 75.1% |
| HellaSwag | — | 95.3% |
| WinoGrande | — | 87.5% |

## Math

- DeepSeek V4 Pro: 64.8 (#30)
- GPT-4: 10.8 (#309)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 1.1% |
| LMArena Math | 1455 | 1269 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |
| MATH Level 5 | — | 23% |
| GSM8K | — | 92% |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#31)
- GPT-4: 18.4 (#282)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| GPQA Diamond | 91.7% | 35.7% |
| LMArena Expert | 1464 | 1211 |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| MMLU | — | 86.4% |
| TriviaQA | — | 84.8% |

## Multilingual

- DeepSeek V4 Pro: 54.4 (#45)
- GPT-4: 40.6 (#215)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| LMArena Non-English | 1439 | 1246 |
| LMArena Chinese | 1486 | 1242 |
| LMArena French | 1472 | 1283 |
| LMArena German | 1458 | 1251 |
| LMArena Japanese | 1445 | 1209 |
| LMArena Korean | 1447 | 1184 |
| LMArena Russian | 1453 | 1251 |
| LMArena Spanish | 1458 | 1261 |

## Instruction Following

- DeepSeek V4 Pro: 76.1 (#47)
- GPT-4: 65.3 (#222)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1241 |

## Long Context

- DeepSeek V4 Pro: 45.0 (#51)
- GPT-4: 37.7 (#212)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| LMArena Longer Query | 1458 | 1244 |
| CL-bench Life | 13.5% | — |

## Writing & Preference

- DeepSeek V4 Pro: 65.5 (#46)
- GPT-4: 34.9 (#268)

| Benchmark | DeepSeek V4 Pro | GPT-4 |
|---|---|---|
| LMArena Text | 1451 | 1263 |
| LMArena Creative Writing | 1446 | 1244 |
| EQ-Bench Creative Writing | 1553 | 752 |
| LMArena Multi-Turn | 1467 | 1257 |
| EQ-Bench 4 | 1166 | — |

## FAQ

### Is DeepSeek V4 Pro better than GPT-4?

DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 29.1 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Pro or GPT-4?

DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; GPT-4 lists at $30 and $60.

### Is DeepSeek V4 Pro or GPT-4 better for coding?

DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 31.6 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Pro does, with 1M tokens against 8K.

### How many benchmarks do DeepSeek V4 Pro and GPT-4 share?

27 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and GPT-4 has 38.