Anthropic canceled OpenAI access to his family Claude from AI models, according to Report in Wired.
Sources informed Wired that Opeli connects Claude with internal tools, which allowed the company to compare Claude performance with its own models in terms corresponding to coding, writing and security.
In the statement given for TechCrunch, Antropic spokesman said: “Openai technical staff also used our coding tools before starting the GPT-5”, in fact “direct violation of our service conditions”. (Anthropic business conditions prohibit corporations to use Claude to build competitive services.)
However, the company also stated that it could proceed to give OpenAI access to “for comparative tests and security purposes”.
Meanwhile, his own statement, OpenAI spokesman, described his use as a “industry standard” and added: “Although we respect Antropica’s decision to cut off our access to the API interface, he is disappointing, given that our API remains available to them.”
Anthropian directors have already shown resistance before providing access to competitors, and the scientific director Jared Kaplan previously justified the company’s decision to cut off Windsurf (rumors about the acquisition of OpenAI, then acquired by Cognition), saying: “I think it would be strange that we would sell Claude to Openai.”