Wei Li

ByteDance's Robix, a unified vision-language model, outperforms GPT-4o and Gemini 2.5 Pro in interactive robot task execution through chain-of-thought reasoning and hierarchical control architecture.

“Co-author of the Robix research paper”