Model comparison

DeepSeek-V3.1 vs GPT-5.2 Pro

GPT-5.2 Pro is the stronger model overall, scoring 52.3 to 42.8 on the Noometry Index. DeepSeek-V3.1 costs 136× less per token, which makes it the better buy when GPT-5.2 Pro's lead doesn't matter for your workload.

Last verified October 10, 2026. 2 shared benchmarks.

DeepSeek-V3.1

DeepSeek

42.8

Rank #108 Confirmed

GPT-5.2 Pro OpenAI

52.3

Rank #39 Reported

Summary

They share 2 benchmarks with published results for both. DeepSeek-V3.1 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 65.3 to 38.9.
The biggest single-benchmark swing is SimpleBench: 40% for DeepSeek-V3.1 and 57.4% for GPT-5.2 Pro.
DeepSeek-V3.1 is cheaper at $0.25 / $0.95 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-V3.1 has downloadable open weights; the other is API-only.

Side by side

DeepSeek-V3.1 and GPT-5.2 Pro specifications
	DeepSeek-V3.1	GPT-5.2 Pro
Provider	DeepSeek	OpenAI
Noometry Index	42.8	52.3
Released	2025-08-21	2025-12-11
Weights	Open	Proprietary
Context window	164K	400K
Max output	8K	128K
Input $ / M tokens	$0.25	$21
Output $ / M tokens	$0.95	$168
Results tracked	27	8

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Not comparable

DeepSeek-V3.1: 40.3 (#144), GPT-5.2 Pro: —

Coding benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
WeirdML	38.4%	—
LMArena Coding	1417	—

Reasoning GPT-5.2 Pro leads

DeepSeek-V3.1: 27.9 (#110), GPT-5.2 Pro: 51.5 (#33)

Reasoning benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
SimpleBench	40%	57.4%
Epoch Capabilities Index	139.92	155.4
ARC-AGI-2	—	54.2%
Kagi LLM Benchmark	53.2%	—
NYT Connections (extended)	—	79.3%
ARC-AGI-1	—	90.5%
LMArena Hard Prompts	1417	—
DTBench	82.7%	—
LMCA	24.3%	—
ForecastBench	58	—

Math GPT-5.2 Pro leads

DeepSeek-V3.1: 38.9 (#122), GPT-5.2 Pro: 65.3 (#29)

Math benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
FrontierMath (Tiers 1-3)	—	74%
FrontierMath Tier 4	—	46%
LMArena Math	1420	—
FrontierMath Tier 4 (v1)	—	31.3%

Knowledge Not comparable

DeepSeek-V3.1: 43.7 (#90), GPT-5.2 Pro: —

Knowledge benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
Vectara Hallucination Rate	5.5%	—
LMArena Expert	1405	—

Multilingual Not comparable

DeepSeek-V3.1: 51.6 (#106), GPT-5.2 Pro: —

Multilingual benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
LMArena Non-English	1400	—
LMArena Chinese	1469	—
LMArena French	1447	—
LMArena German	1411	—
LMArena Japanese	1378	—
LMArena Korean	1337	—
LMArena Russian	1405	—
LMArena Spanish	1431	—

Instruction Following Not comparable

DeepSeek-V3.1: 73.9 (#110), GPT-5.2 Pro: —

Instruction Following benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
LMArena Instruction Following	1400	—

Long Context Not comparable

DeepSeek-V3.1: 36.3 (#232), GPT-5.2 Pro: —

Long Context benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
Fiction.LiveBench	52.8%	—
LMArena Longer Query	1422	—

Writing & Preference Not comparable

DeepSeek-V3.1: 60.3 (#98), GPT-5.2 Pro: —

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.1	GPT-5.2 Pro
LMArena Text	1420	—
LMArena Creative Writing	1401	—
EQ-Bench Creative Writing	1436	—
LMArena Multi-Turn	1408	—

Frequently asked questions

Is DeepSeek-V3.1 better than GPT-5.2 Pro?

Which is cheaper, DeepSeek-V3.1 or GPT-5.2 Pro?

DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 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-V3.1 and GPT-5.2 Pro share?

2 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and GPT-5.2 Pro has 8.