Anthropic revokes OpenAI’s access to Claude models
Anthropic has recently decided to cut off OpenAI’s access to its Claude family of AI models, as reported by Wired. This move marks a significant shift in the competitive landscape among leading artificial intelligence companies.
Reasons behind the decision
According to internal sources, OpenAI was using Claude to compare performance with its own models in areas such as writing, coding, and safety. Anthropic stated that OpenAI’s technical team used Claude-based coding tools ahead of the GPT-5 launch, violating commercial terms that prohibit using Claude to build competing services.
"OpenAI’s own technical staff were also using our coding tools ahead of the launch of GPT-5, which is apparently a direct violation of our terms of service."
Anthropic spokesperson
Limited access for benchmarking and safety
Despite the cutoff, Anthropic confirmed that OpenAI will still have access to Claude models solely for benchmarking and safety evaluations, maintaining minimal collaboration to ensure transparency and safety in the industry.
OpenAI’s response
OpenAI described its use of Claude models as “industry standard” and expressed disappointment at Anthropic’s decision, highlighting that its own API remains available to Anthropic.
"While we respect Anthropic’s decision to cut off our API access, it’s disappointing considering our API remains available to them."
OpenAI spokesperson
Implications for the AI sector
- Competition among AI companies is intensifying, with increasing restrictions on access to proprietary models.
- Companies are adopting stricter policies to protect their technology and competitive advantage.
- Collaboration between industry leaders remains crucial for safety and benchmarking, but is becoming more limited.