744B / 40B (MoE)
MIT
Flagship MoE model trained entirely on Huawei Ascend 910 chips with MindSpore framework; 256 experts with 8 activated per token; Dynamically Sparse Attention (DSA) mechanism using DeepSeek Sparse Attention; 200K context window with 131K max output; agentic intelligence mode for long-running tasks; specialized in creative writing, coding, advanced reasoning, and long-context processing; #1 open-source model globally at release; strong benchmarks: SWE-bench Verified (77.8%), Humanity's Last Exam (50.4), τ²-Bench (89.7), BrowseComp (75.9); approaches Claude Opus 4.5 in coding benchmarks; surpasses Gemini 3 Pro on some benchmarks; supports data insights, smart writing, full-stack development; 17-19 tokens/sec throughput; zero dependency on US-manufactured hardware