Anthropic has released its large model series Claude 3, where Opus surpasses OpenAI's GPT-4 in evaluation metrics and visual recognition capabilities. Anthropic's team focuses on the safety research of large models and has developed "Constitutional AI" to address behavioral issues of models. They have successfully raised over $7 billion, providing ample resources for the development and training of large models. In the tech field, Anthropic's Opus model has garnered widespread attention, with a research focus on making large models safer and exploring the concept of "Constitutional AI".