DeepSeek: DeepSeek V3.2

By DeepSeek

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...

OpenRouter rank

0.02% share

Context window

131K tokens

Input price

$0.25 / 1M tok

Output price

$0.38 / 1M tok

Modalities

text

External ID

deepseek/deepseek-v3.2

No ranking history yet

This model is in our catalog but hasn't appeared on a tracked leaderboard yet. Trend lines appear after the next ingestion cycle (within 6 hours).

Back to all rankings