Apa yang dijawab halaman ini
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 att
- DeepSeek · deepseek/deepseek-v3.2
- text->text · rute model China
- 163.840 context · US$0,26 input