Plain-English Summary
An open DeepSeek model that combines general chat, reasoning, and tool use.
DeepSeek-V3.2
An open DeepSeek model that combines general chat, reasoning, and tool use.
DeepSeek V3
Reasoning LLM
Available / older
Open weights: Yes
API: Yes
Technical Notes
MoE language model integrating thinking and tool-use behavior with long-context support.
Best For
Open deployment, coding, reasoning, research
Watch Out For
Older than V4; deployment and prompt behavior can be less polished than hosted frontier models.
Model Specs
- Maker
- DeepSeek
- Country
- China
- Family
- DeepSeek V3
- Type
- Reasoning LLM
- Release date
- 2025-12-01
- Date confidence
- Exact
- Status
- Available / older
- Context window
- 128,000 tokens
- Max output
- 64,000
- Parameters
- Not disclosed
- Active parameters
- Not disclosed
- Architecture
- Mixture-of-experts transformer
- Reasoning
- Yes
- Tool calling
- Yes
- Structured output
- Yes
- API available
- Yes
- Open weights
- Yes
- Self-hostable
- Yes
- License
- DeepSeek model license
- Inputs
- Text
- Outputs
- Text
- Approx. price
- See provider pricing
- How to access
- Provider website or API
- Model / API ID
- deepseek-v3.2
- Knowledge cutoff
- Not publicly disclosed
- Fine-tuning
- Fine-tuning and self-hosting
- Completeness
- 90 %
- Verified
- Jul 18, 2026
- Snapshot date
- Jul 18, 2026
- Source
- Official source