DeepSeek: DeepSeek V3.2
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...
Everything else
| Max output | 66K |
|---|---|
| Input / 1M | $0.269 |
| Output / 1M | $0.4 |
| Cache read / 1M | $0.135 |
| Cache write / 1M | — |
| 1M in + 1M out | $0.669 |
| Takes | text |
| Returns | text |
| Tools | yes |
| Reasoning | optional |
| Open weights | yes |
| Released | 2025-12-01 |
| Knowledge cutoff | — |
| Retires | — |
| Catalog id | deepseek/deepseek-v3.2 |
Benchmarks
| source | epoch |
|---|---|
| eci | 146.59 |
| confidence | Likely |
| accessibility | Open weights (unrestricted) |
Benchmark scores by Epoch AI, CC BY 4.0.
Using it
In 00, this model is picked per agent — and per tier, so the model that answers your customers need not be the one that writes your code. Get 00.