DEDeepSeek/ DeepSeek
DeepSeek V3.2 Exp
deepseek/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...
Model information
Released
2025-09-29
Context window
164K
Max output
66K
Model parameters
685B
Weights
Open
Modalities
Input
Text
Output
Text
Identity
- Version
- deepseek-v3.2-exp
- Knowledge cutoff
- 2025-07-31
- License
- MIT
Resources
Capabilities
Open weightReasoningStructuredTemperatureTool call
Providers
| Provider | Provider model ID | Protocol | Context | Max output | Input | Output | Capabilities | Notes | Pricing | Streaming |
|---|---|---|---|---|---|---|---|---|---|---|
| OpenRouter | deepseek/deepseek-v3.2-exp | openai_chat | 164K | 66K | ReasoningTool callStructuredTemperature | Hugging Face: deepseek-ai/DeepSeek-V3.2-Exp | input $0.27/M · output $0.41/M | Yes | ||
| ZenMux | deepseek/deepseek-v3.2-exp | openai_chat | 164K | 66K | ReasoningTool callTemperature | - | input $0.22/M · output $0.33/M | Yes | ||
| Kilo Gateway | deepseek/deepseek-v3.2-exp | openai_chat | 164K | 66K | ReasoningTool callStructuredTemperature | - | input $0.27/M · output $0.41/M | Yes | ||
| NovitaAI | deepseek/deepseek-v3.2-exp | openai_chat | 164K | 66K | ReasoningTool callStructuredTemperature | - | input $0.27/M · output $0.41/M | Yes | ||
| Qiniu | deepseek/deepseek-v3.2-exp | openai_chat | 164K | 66K | Tool callTemperature | - | input ¥0.002/K · output ¥0.003/K | Yes | ||
| DeepSeek | deepseek-v3.2-exp | custom | 164K | 66K | - | - | - | Yes |