Neura Intelligence Index

AI Model Benchmarks

DeepSeek R1 Zero vs DeepSeek-V3.2-Exp — Benchmark Comparison | Neura Market

Back to Leaderboard

DeepSeek R1 Zero vs DeepSeek-V3.2-Exp

DeepSeek R1 Zero

DeepSeek 🇨🇳

63.0

#99—

DeepSeek-V3.2-Exp

DeepSeek 🇨🇳

63.5

#96—

Category Comparison

Benchmark-by-Benchmark

reasoning

Benchmark	DeepSeek R1 Zero	DeepSeek-V3.2-Exp
GPQA Diamond	73.3	79.9
Humanity's Last Exam	—	19.8

math

Benchmark	DeepSeek R1 Zero	DeepSeek-V3.2-Exp
AIME 2025	—	89.3

coding

Benchmark	DeepSeek R1 Zero	DeepSeek-V3.2-Exp
SWE-bench Verified	—	67.8
TerminalBench	—	37.7

general

Benchmark	DeepSeek R1 Zero	DeepSeek-V3.2-Exp
SimpleQA	—	97.1

agent

Benchmark	DeepSeek R1 Zero	DeepSeek-V3.2-Exp
BrowseComp	—	40.1

Category Winners

coding

DeepSeek-V3.2-Exp

58.6

math

DeepSeek-V3.2-Exp

68.6

reasoning

DeepSeek R1 Zero

63.0

general

DeepSeek-V3.2-Exp

98.8

agent

DeepSeek-V3.2-Exp

11.3