#26
DeepSeek V3.2
deepseek-v3.2-exp
0.0
Helpfulness
Instruction Following
Comprehension
Empathy
Creative Writing
Helpfulness
0.0
Instruction Following
0.0
Comprehension
0.0
Empathy
0.0
Creative Writing
0.0
Speed
Avg 95 tok/s
Release Date
September 29, 2025
Lab
DeepSeek
Type
Open Source
Context Size
160K
Max Output Tokens
8.2K
Cost per 1 million tokens
$0.07 / $0.14
Model Inputs*
Text
Model Outputs*
Text
Tool Calling*
Enabled
Overall Assistant Score
An average score combining the 5 main categories.
81.95 pts
Rank #26
48th Percentile
0
Novice
33
Capable
66
Proficient
100
Expert
DeepSeek V3.2 Exp is an experimental model introducing the new DeepSeek Sparse Attention (DSA) mechanism for superior long-context efficiency. Built on the V3.1-Terminus architecture, it delivers high-performance inference at significantly reduced costs. This model excels in general-purpose tasks and coding, offering a massive 160k context window while maintaining competitive benchmark scores against proprietary alternatives.

Intelligence

Overall Score; Higher is better

GPT-5 Mini
o4-mini
Grok 4 Fast
o1
DeepSeek V3…
DeepSeek V3…
o3-mini
Grok 3
Grok 4
DeepSeek R1

Speed

Output Tokens per Second; Higher is better

GPT-5 Nano
Llama 3.3 8…
Llama 4 Sco…
Gemini 2.5 …
o4-mini
DeepSeek V3…
Llama 4 Mav…
GPT-5 Mini
Gemini 2.0 …
Claude Haik…

Price

USD per 1M Tokens; Lower is better

GLM 4.5 Air
GPT-5 Nano
Llama 3.3 8…
DeepSeek V3…
DeepSeek V3…
DeepSeek V3…
GLM 4.7 Fla…
Gemini 2.0 …
Gemini 2.0 …
Gemini 2.5 …

deepseek Models

Overall Score; Same provider comparison

DeepSeek V3…
DeepSeek V3…
DeepSeek V3…
DeepSeek R1
DeepSeek V3…
DeepSeek V3

DeepSeek Family

Overall Score; Same model family comparison

DeepSeek V3…
DeepSeek V3…
DeepSeek V3…
DeepSeek R1
DeepSeek V3…
DeepSeek V3

Closest Rivals

Overall Score; Nearest by overall rank

GPT-5 Mini
o4-mini
Grok 4 Fast
o1
DeepSeek V3…
DeepSeek V3…
o3-mini
Grok 3
Grok 4
DeepSeek R1