Hajra Naz
Chinese artificial intelligence startup DeepSeek unveiled an upgrade to its flagship V3 model on Thursday, introducing faster processing speeds and new optimization features designed for Chinese-made chips.
The move underscores China’s efforts to strengthen its domestic semiconductor ecosystem, as Beijing pushes to reduce reliance on U.S. technology amid ongoing export restrictions from Washington.
DeepSeek has already made global headlines this year by releasing AI models that rival Western systems such as OpenAI’s ChatGPT while maintaining lower operational costs.
The new V3.1 update marks the company’s third model upgrade in 2025, following the R1 update in May and an earlier V3 enhancement in March.
In a WeChat post, DeepSeek said the V3.1 model supports the UE8M0 FP8 precision format, specifically optimized for “soon-to-be-released next-generation domestic chips.” However, the company did not disclose which chipmakers or models would be supported.
Read More: Deepseek AI Enhances Data Capacity and Makes AI More Reachable
The FP8 (8-bit floating point) format enables AI systems to run more efficiently by consuming less memory while maintaining high speeds—an increasingly important capability for large-scale generative AI models.
According to the company, the upgraded model features a hybrid inference structure that can operate in both reasoning and non-reasoning modes. Users can switch between the two via a new “deep thinking” button on DeepSeek’s app and web platform, both of which now run the V3.1 version.
DeepSeek also announced that starting September 6, it will adjust pricing for developers using its API platform, which allows third-party applications and web products to integrate DeepSeek’s AI models.
By aligning its technology with China’s domestic chip ecosystem, DeepSeek appears to be positioning itself as a key player in the country’s long-term strategy to challenge U.S. dominance in semiconductors and artificial intelligence.
