DeepSeek V4 Released: Open-Source LLM Challenges Closed-Source Leaders and Partners with Huawei Chips
DeepSeek V4 launches in two versions—Pro and Flash—offering 1 M token context, enhanced agent capabilities, world‑knowledge and reasoning performance, a new token‑compression attention mechanism with DSA sparse attention, Huawei compute support, updated APIs, and a migration plan for legacy models.
