Anthropic blocked OpenAI from accessing Claude due to policy violations
Anthropic has suspended OpenAI’s access to the API for its AI model Claude, citing alleged use of the tool to create a competing product, which violates its terms of service.
Anthropic has suspended OpenAI’s access to the API for its AI model Claude, citing alleged use of the tool to create a competing product, which violates its terms of service.
As WIRED reports, on July 30, Anthropic officially disconnected OpenAI from the Claude model API. Sources claim that OpenAI technicians were actively using Claude Code for internal tasks before the launch of GPT-5. This violated the terms of service, which explicitly prohibit using Claude to develop a competing product, train other models, or reverse engineer it.
According to Anthropic spokesperson Christopher Nulty, OpenAI has been using Claude for tests related to coding, creative tasks, and security queries, including malicious content. OpenAI connected Claude not through a regular interface, but directly through an API. This allowed for extensive comparisons with its own models, including GPT-5.
OpenAI commented on the situation with disappointment. OpenAI representative Anna Wong stated that «evaluating other models is standard practice in the industry,» and also reminded that OpenAI’s API remains open to Anthropic. At the same time, Anthropic added that they remain open to «safe benchmarking,» but clearly distinguish it from potential copying of Claude’s functions.
This is not the first time Anthropic has disabled access to its models: similar sanctions were previously applied to the startup Windsurf, which was rumored to be acquired by OpenAI. Such practices are not uncommon in the technology sector, with Facebook and Salesforce previously taking similar actions.
Claude is one of the leading AI models from Anthropic, which is actively promoted as an alternative to ChatGPT. The company was founded by former OpenAI engineers and is one of the main competitors in the generative AI market. Losing access to Claude could complicate testing and benchmarking new OpenAI products, especially in the context of the expected release of GPT-5.
We previously wrote about how Anthropic announced that it was introducing new weekly limits on the use of Claude AI to limit paid subscribers who use the Claude Code artificial intelligence programming tool 24/7.
Anthropic burned millions of books to train its AI: court finds it legal
Anthropic burned millions of books to train its AI: court finds it legal
«Anthropic positions itself as the white knight of the AI industry, but it is not.» Why Reddit sued the AI company
«Anthropic positions itself as the white knight of the AI industry, but it is not.» Why Reddit sued the AI company
Anthropic’s Claude Opus 4 AI model tries to blackmail developers in test scenarios
Anthropic’s Claude Opus 4 AI model tries to blackmail developers in test scenarios
Read the country’s main IT news in our Telegram
Read the country’s main IT news in our Telegram
https://feedback.challonge.com/forums/44455-feature-requests/suggestions/50384145-the-critical-importance-of-security-checks-when-hi
https://www.freecodecamp.org/news/learn-key-system-design-principles-behind-high-traffic-platforms-like-gaming-and-job-discovery/
Hiring AI Engineers
https://www.freecodecamp.org/news/learn-key-system-design-principles-behind-high-traffic-platforms-like-gaming-and-job-discovery/
https://feedback.challonge.com/forums/44455-feature-requests/suggestions/50375346-europe-s-data-vision-dataspaces-for-zero-trust-ai
https://www.freecodecamp.org/news/learn-key-system-design-principles-behind-high-traffic-platforms-like-gaming-and-job-discovery
https://hackernoon.com/europes-data-vision-dataspaces-for-zero-trust-ai-infrastructure
Hire Solana Developers
https://upstaff.com/hire/solana/