Anthropic revokes OpenAI’s access to Claude models over competitive concerns

Anthropic has cut off OpenAI’s access to its Claude family of AI models, according to a report by Wired. The move comes after sources revealed that OpenAI had been using Claude’s APIs to evaluate its own models across key areas such as coding, writing, and safety, likely in preparation for the upcoming launch of GPT-5. In a statement to Wired, an Anthropic spokesperson confirmed the revocation, citing a breach of the company’s terms of service. The spokesperson noted that OpenAI’s technical staff had been using Anthropic’s coding tools ahead of the GPT-5 launch, which violates the commercial agreement that prohibits customers from using Claude to develop competing AI services. OpenAI responded by expressing disappointment, stating, “While we respect Anthropic’s decision to cut off our API access, it’s disappointing considering our API remains available to them.” The statement underscores the growing tension between leading AI companies over data access and competitive benchmarking. This development reflects a broader shift in the AI industry, where companies are increasingly guarding their models and tools as strategic assets. Anthropic’s leadership has long emphasized caution around sharing access with rivals. Chief Science Officer Jared Kaplan previously defended similar decisions, including cutting off Windsurf—a company reportedly in talks with OpenAI—by saying it would be “odd” for Anthropic to provide Claude access to OpenAI, especially given the competitive landscape. As the race to build the most advanced AI systems intensifies, access to high-performing models is becoming a critical battleground. With OpenAI now blocked from using Claude for internal evaluation, the competitive dynamic between the two companies is likely to grow even more pronounced.