DeepSeek V4 launch, officially partners with Huawei chips

DeepSeek's official WeChat public account announced that DeepSeek-V4 possesses a million-word ultra-long context, achieving a leading position in both domestic and open-source fields in terms of agent capabilities, world knowledge, and reasoning performance.

Currently, DeepSeek-V4 has become the Agentic Coding model used by the company's internal employees. According to evaluation feedback, the user experience is superior to Sonnet 4.5, and the delivery quality is close to the non-thinking mode of Opus 4.6, but there is still a certain gap with the thinking mode of Opus 4.6.

DeepSeek-V4-Pro: Performance Rivaling Top Closed-Source Models

DeepSeek-V4 has pioneered a new attention mechanism that compresses on the token dimension, combined with DSA sparse attention (DeepSeek Sparse Attention), achieving a globally leading long-context capability, and significantly reducing the demand for computation and memory compared to traditional methods. From now on, 1M (one million) context will be the standard for all official DeepSeek services.

First Adaptation to Domestic Chips, Supply Chain Signaling Significant