Anthropic cuts off OpenAI’s access to its Claude models
Anthropic has revoked OpenAI’s access to its Claude AI models, alleging OpenAI breached terms by using Claude for benchmarking and to build competing services ahead of GPT-5. While Anthropic will still allow limited API use for safety evaluations, it accuses OpenAI of misusing coding tools. OpenAI calls the move "disappointing" and insists its usage was "industry standard."
Also mentioned in:
- TechCrunch — OpenAI and Anthropic Open Models for Cross-Lab Safety Testing, Urge Routine Collaboration