Read, Watch or Listen

Anthropic revoked OpenAI's API access to its Claude models due to OpenAI violating terms of service by using Claude to potentially develop GPT-5. OpenAI reportedly used Claude for internal benchmarking and safety testing, comparing its capabilities to their own models. Anthropic claims this breaches its terms prohibiting using its service to build competing products. OpenAI says this is standard industry practice for evaluating AI systems, while Anthropic maintains it will allow access for benchmarking and safety. This follows a similar action against Windsurf, another AI startup.

Prepared by Jonathan Pierce and reviewed by editorial team.