In a bold move that’s shaking up the tech world, Anthropic has cut off OpenAI’s access to its Claude AI platform, citing terms of service violations. The decision comes at a critical moment as OpenAI prepares to launch its highly anticipated GPT-5 model.

According to sources, OpenAI was using special developer APIs to test Claude’s capabilities in various domains, including coding, creative writing, and safety evaluations. Anthropic claims this usage directly violated their commercial terms of service, which prohibit customers from building competing products or reverse engineering their technology.

Christopher Nulty, an Anthropic spokesperson, stated that Claude Code has become incredibly popular among developers, and OpenAI’s actions crossed a line. “Unfortunately, this is a direct violation of our terms of service,” Nulty explained.

OpenAI’s chief communications officer Hannah Wong responded diplomatically, acknowledging that evaluating other AI systems is an industry standard practice for benchmarking and improving safety. “While we respect Anthropic’s decision to cut off our API access, it’s disappointing,” Wong said.

This isn’t the first time Anthropic has restricted API access. Last month, they similarly blocked AI coding startup Windsurf after rumors of a potential OpenAI acquisition. Jared Kaplan, Anthropic’s chief science officer, previously commented that selling Claude to OpenAI would be “odd”.

The incident highlights the increasingly competitive and territorial landscape of artificial intelligence development. Top tech companies have long used API access restrictions as a strategic tool, with precedents like Facebook’s actions against Vine and Salesforce’s recent data API limitations.

As the AI arms race continues, this move by Anthropic signals a growing tension between innovation and protecting proprietary technology. With GPT-5 on the horizon and Claude’s rising popularity, the stakes have never been higher in the world of artificial intelligence.

While Anthropic suggests they will continue to allow API access for benchmarking and safety evaluations, the details remain unclear. The tech community will be watching closely to see how this drama unfolds.