Anthropic has restricted OpenAI’s access to its Claude AI models
By ForGeeks
As reported by Wired, Anthropic has revoked OpenAI’s access to its Claude family of AI models.
According to the publication’s sources, OpenAI had been using Claude to connect to internal tools to compare its performance with its own models in areas such as programming, text generation and security.
In a statement to TechCrunch, an Anthropic spokesperson noted:”OpenAI technicians also used our development tools prior to the GPT-5 release, which appears to be a direct violation of our terms of use.” Indeed, Anthropic’s commercial policy prohibits companies from using Claude to create competing services.
The company clarified, however, that it will retain OpenAI’s access to its models “for benchmarking and security assessment”.
In response, an OpenAI spokesperson called such practices “industry standard” and added:”We respect Anthropic’s decision to limit access to our API, but find this a disappointing move, especially given that our API is still available to them.”
Anthropic executives have previously expressed their displeasure with granting access to competitors. For example, the company’s chief scientist Jared Kaplan explained the rejection of a deal with Windsurf (a startup that OpenAI was rumored to be interested in but was eventually bought by Cognition) as follows:”It would be weird if we sold to Claude OpenAI.”.
The Anthropic has restricted OpenAI’s access to its Claude AI models was first published on itzine.ru.
