# DeepSeek-V3 vs GPT-5.2 Pro

> GPT-5.2 Pro is the stronger model overall, scoring 52.2 to 39.2 on the Noometry Index. DeepSeek-V3 costs 143× less per token, which makes it the better buy when GPT-5.2 Pro's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-v3-vs-gpt-5-2-pro
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and GPT-5.2 Pro in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.2 Pro leads 64.8 to 30.4.
- The biggest single-benchmark swing is SimpleBench: 27.2% for DeepSeek-V3 and 57.4% for GPT-5.2 Pro.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $21 / $168 for GPT-5.2 Pro.
- GPT-5.2 Pro accepts more context: 400K tokens versus 164K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 39.2 | 52.2 |
| Rank | 157 | 34 |
| Context | 164K | 400K |
| Input $/M | $0.24 | $21 |
| Output $/M | $0.90 | $168 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek-V3: 41.9 (#99)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| Aider Polyglot | 55.1% | — |
| SciCode | 35.8% | — |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| LMArena Coding | 1368 | — |
| BigCodeBench Complete | 62.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |

## Agentic & Tool Use

- DeepSeek-V3: —
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| METR Time Horizons | 49.6% | — |

## Reasoning

- DeepSeek-V3: 20.0 (#226)
- GPT-5.2 Pro: 51.8 (#31)

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| SimpleBench | 27.2% | 57.4% |
| Epoch Capabilities Index | 135.95 | 155.41 |
| ARC-AGI-2 | — | 54.2% |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 79.3% |
| ARC-AGI-1 | — | 90.5% |
| CritPt | 0% | — |
| LiveBench Reasoning | 65.8% | — |
| LMArena Hard Prompts | 1365 | — |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |

## Math

- DeepSeek-V3: 30.4 (#208)
- GPT-5.2 Pro: 64.8 (#23)

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 74% |
| FrontierMath Tier 4 | — | 46% |
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| LMArena Math | 1373 | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
| FrontierMath Tier 4 (v1) | — | 31.3% |

## Knowledge

- DeepSeek-V3: 37.5 (#147)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| GPQA Diamond | 67.6% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| LMArena Expert | 1352 | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |

## Multilingual

- DeepSeek-V3: 48.8 (#138)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| LMArena Non-English | 1358 | — |
| LMArena Chinese | 1390 | — |
| LMArena French | 1386 | — |
| LMArena German | 1374 | — |
| LMArena Japanese | 1333 | — |
| LMArena Korean | 1321 | — |
| LMArena Russian | 1372 | — |
| LMArena Spanish | 1358 | — |

## Instruction Following

- DeepSeek-V3: 73.0 (#125)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
| LMArena Instruction Following | 1345 | — |

## Long Context

- DeepSeek-V3: 34.0 (#248)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| Fiction.LiveBench | 50% | — |
| LMArena Longer Query | 1353 | — |

## Writing & Preference

- DeepSeek-V3: 57.6 (#124)
- GPT-5.2 Pro: —

| Benchmark | DeepSeek-V3 | GPT-5.2 Pro |
|---|---|---|
| LMArena Text | 1375 | — |
| LMArena Creative Writing | 1364 | — |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LMArena Multi-Turn | 1389 | — |
| LiveBench Language | 49.1% | — |

## FAQ

### Is DeepSeek-V3 better than GPT-5.2 Pro?

GPT-5.2 Pro is the stronger model overall, scoring 52.2 to 39.2 on the Noometry Index. DeepSeek-V3 costs 143× less per token, which makes it the better buy when GPT-5.2 Pro's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek-V3 or GPT-5.2 Pro?

DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; GPT-5.2 Pro lists at $21 and $168.

### Which has the bigger context window?

GPT-5.2 Pro does, with 400K tokens against 164K.

### How many benchmarks do DeepSeek-V3 and GPT-5.2 Pro share?

2 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and GPT-5.2 Pro has 8.