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 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.

ChatDeepSeek164k tokens$0.2943 / $0.4469 · 1M

deepseek/deepseek-v3.2-exp

Context
164k tokens
Completion cap
65,536
Tools
Yes
JSON
Yes
Released
2025-09-29

Call it from Geek Hub

Same OpenAI SDK. Change the base URL and the model id.

curl https://api.geekhub.mx/v1/chat/completions \
  -H "Authorization: Bearer $GEEKHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v3.2-exp",
    "messages": [{"role": "user", "content": "Hola"}]
  }'
Get an API key

Frequently asked questions

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 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 runs on the Geek Hub API (OpenAI-compatible). Model id: deepseek/deepseek-v3.2-exp.
Is DeepSeek V3.2 Exp free?
No. Input is $0.2943 per 1M tokens and output is $0.4469 per 1M tokens on Geek Hub (markup included).
What is the context length of DeepSeek V3.2 Exp?
DeepSeek V3.2 Exp has a 164k tokens context window. It supports up to 65,536 completion tokens.
Does DeepSeek V3.2 Exp support tool calling and structured outputs?
DeepSeek V3.2 Exp accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
When was DeepSeek V3.2 Exp released?
DeepSeek V3.2 Exp was released on 2025-09-29.

More models from DeepSeek