EN Submit a tool

DeepSeek V3.2 Exp

Rank #118 deepseek

deepseek/deepseek-v3.2-exp

Usage data date 2026-07-26 · updated daily

Usage and specs

Daily rank#118
Daily tokens16.7B
Daily requests2.6M
Daily change −3.3%
Context window131.1K
Released

What is DeepSeek V3.2 Exp

DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism designed to improve training and inference efficiency in long-context scenarios while maintaining output quality. Users can control the reasoning behaviour with the `reasoning` `enabled` boolean. [Learn more in our docs](https://openrouter.ai/docs/use-cases/reasoning-tokens#enable-reasoning-with-default-config) The model was trained under conditions aligned with V3.1-Terminus to enable direct comparison. Benchmarking shows performance roughly on par with V3.1 across reasoning, coding, and agentic tool-use tasks, with minor tradeoffs and gains depending on the domain. This release focuses on validating architectural optimizations for extended context lengths rather than advancing raw task accuracy, making it primarily a research-oriented model for exploring efficient transformer designs.

DeepSeek V3.2 Exp capabilities

Input: Text   Output: Text

Reasoning output Reasoning JSON output Structured outputs Tool choice Tool use Knowledge cutoff 2025-07-31

DeepSeek V3.2 Exp API pricing

Input · per 1M tokens$0.27
Output · per 1M tokens$0.41

List prices in USD; see each provider’s site for current terms.