Anthropic has revoked API access to its Claude family of AI models for competitor OpenAI, citing a “direct violation of our terms of service” after OpenAI used Claude for performance comparisons against its own models.
In a statement to TechCrunch, an Anthropic spokesperson explained that OpenAI’s technical staff were using Claude’s coding tools ahead of the launch of GPT-5, which is prohibited under Anthropic’s terms of service as it involves using Claude to develop competing services. However, Anthropic clarified that it would continue to permit access specifically for the purposes of benchmarking and safety evaluations, indicating a nuanced approach to its API access policies.
OpenAI responded to the decision by describing its usage of Claude as “industry standard” and expressed disappointment with Anthropic’s move, highlighting that OpenAI’s own API remains available to Anthropic. This exchange underscores the competitive dynamics between the two AI companies.
Anthropic’s action is consistent with its previous stance on providing services to competitors, as evidenced by its Chief Science Officer, Jared Kaplan’s, explanation for cutting off access to the startup Windsurf due to its status as an OpenAI acquisition target. This pattern suggests that Anthropic is vigilant about its API access and is willing to enforce its terms of service to protect its interests.
