You can now use DeepSeek's V4 models in OpenClaw
What's the story
OpenClaw, a popular artificial intelligence agent, has integrated DeepSeek's latest V4 Flash model as its default. The move comes amid global scrutiny over the implications of the Chinese AI company's major update, which is optimized for Huawei's chips. The integration was announced today and also included other features like adding Google Meet into the agent.
Usage
OpenClaw's update follows DeepSeek's V4 model launch
The update from OpenClaw comes just two days after Hangzhou-based DeepSeek unveiled its open-source V4 models. The flagship model, V4 Pro, features a whopping 1.6 trillion parameters, making it the company's largest model to date. The smaller version, V4 Flash, packs an impressive 284 billion parameters. These new models are designed to work with popular agent tools like Anthropic's Claude Code and Tencent Holdings' CodeBuddy.
Collaboration
Huawei's partnership with DeepSeek
The global tech community is closely watching DeepSeek's collaboration with Huawei. The Chinese tech giant has promised "full support" from its Ascend chips and supernode systems for model inference. This partnership is seen as a major test for China's home-grown AI hardware, further highlighting the significance of DeepSeek's V4 models in the global tech landscape.