DeepSeek V3.2 Exp
DeepSeek V3.2 Exp is DeepSeek's language model with a 164K context window and up to 66K output tokens, available from 4 providers, starting at $0.270 / 1M input and $0.410 / 1M output. An experimental DeepSeek V3.2 variant introducing the DeepSeek Sparse Attention mechanism for enhanced long-context processing efficiency.
Capabilities
Input1/5
✓
·
·
·
·
Output1/5
✓
·
·
·
·
Capabilities7/13
✓
·
✓
✓
✓
✓
·
·
·
·
·
✓
✓
Pricing by Provider
| Provider | Standard | Batch | ||
|---|---|---|---|---|
| Input $ / 1M | Output $ / 1M | Input $ / 1M | Output $ / 1M | |
 Hugging Face | $0.270 | $0.410 | — | — |
 Novita | $0.270 | $0.410 | — | — |
 OpenRouter | $0.270 | $0.410 | — | — |
 Alibaba Qwen | $0.287 | $0.431 | $0.143 | $0.215 |
Cost Calculator
Preset:
Compares every provider & tier in USD
Other models
| Model | Tier | Released | Context | Input / 1M | Output / 1M |
|---|---|---|---|---|---|
| DeepSeek V3.2 | — | 164K | $0.252 | $0.378 |