News Summary

Anthropic has decided to revoke OpenAI’s access to its Claude family of AI models, citing a breach of terms of service. The decision was made after it was revealed that OpenAI was allegedly using Claude to compare performance against its own models in areas such as coding, writing, and safety. Despite this restriction, Anthropic has stated it will allow OpenAI access for benchmarking and safety evaluations. This decision comes amidst OpenAI’s ongoing development of GPT-5 and highlights the competitive tensions in the AI industry.