The Model Index
Up to four models

DeepSeek-V3.2-Exp vs DeepSeek-V3.1

On the index (GPT-4 = 100), DeepSeek-V3.2-Exp scores 140 and DeepSeek-V3.1 128. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to DeepSeek-V3.2-Exp
Presets
DeepSeek-V3.2-Exp
140124–157
WeightsOpen
ReleasedSep 2025
Price in / out$0.27 / $0.41
Context164K
Core tests taken0 of 9
DeepSeek-V3.1
128111–145
WeightsOpen
ReleasedAug 2025
Price in / out$0.25 / $0.95
Context164K
Core tests taken0 of 9

DeepSeek-V3.2-Exp leads DeepSeek-V3.1 by 12 points, and costs 1.4× less.

Test by testthird-party results

DeepSeek-V3.2-ExpDeepSeek-V3.1
0%25%50%75%100%REASONING & KNOWLEDGEFiction.LiveBenchFiction.LiveBenchCODINGWeirdMLWeirdMLAGENTSDTBenchDTBenchLMCALMCA
Every result as a table
TestDeepSeek-V3.2-ExpDeepSeek-V3.1
Reasoning & knowledge
Fiction.LiveBench83.3%52.8%
Coding
WeirdML39.5%38.4%
Agents
DTBench87.7%82.7%
LMCA29.1%24.3%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG