ModelWatch
← Catalog

DeepSeek V3.2

DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...

open-weightslow-cost
Input price
$0.26 / 1M tokens
Output price
$0.38 / 1M tokens
Context window
164K tokens
Max output
147K tokens
Output speed
38 tokens/s
Time to first token
3.25s
Inputs
text
Open weights
Yes
OpenRouter id
deepseek/deepseek-v3.2
Listed
Dec 1, 2025

Benchmarks

Intelligence Index
n/a
Engineering Index
n/a

Speed and benchmarks for “DeepSeek V3.2 (Non-reasoning)” from Artificial Analysis. Prices, context window and listing date are live from OpenRouter.

Compare with others