OpenClaw Integrates DeepSeek V4 Models Amid Huawei Chip Scrutiny
OpenClaw has selected DeepSeek's V4 Flash as the standard model for its AI agent, as the tech industry evaluates the effects of DeepSeek's enhancements for Huawei chips. In a Sunday update, OpenClaw broadened its offerings to incorporate both DeepSeek's V4 Flash and its premier V4 Pro models, alongside new Google Meet functionalities. Additionally, OpenClaw has fine-tuned the DeepSeek V4 models to ensure reliability in multi-step tasks. On Friday, DeepSeek revealed that its V4 models are tailored for various agent tools, including OpenClaw, Anthropic's Claude Code, and Tencent's CodeBuddy. Notably, the V4 Pro model boasts 1.6 trillion parameters, and Huawei confirmed complete support for model inference using its Ascend chips and supernode systems, showcasing China's domestic AI technology.
Key facts
- OpenClaw adopted DeepSeek V4 Flash as default model
- OpenClaw expanded catalogue to include V4 Flash and V4 Pro
- OpenClaw integrated Google Meet capabilities
- OpenClaw optimized DeepSeek V4 for multi-step tasks
- DeepSeek V4 models optimized for Claude Code, OpenClaw, CodeBuddy
- DeepSeek V4 Pro has 1.6 trillion parameters
- Huawei announced full support from Ascend chips and supernode systems
- Update on Sunday, DeepSeek announcement on Friday
Entities
Institutions
- OpenClaw
- DeepSeek
- Huawei Technologies
- Anthropic
- Tencent Holdings
Locations
- China