Particle.news
Download on the App Store
4 ARTICLES
·
yesterday

Anthropic Revokes OpenAI’s General Access to Claude API

Anthropic now confines OpenAI’s Claude use to benchmarking and safety evaluations following a terms-of-service breach in GPT-5 development.

Anthropic’s Dario Amodei and Daniela Amodei. Image: Anthropic
Anthropic cuts off OpenAI's Claude access

Overview

  • Anthropic cut off OpenAI’s general Claude API access on July 29 after determining its engineers used Claude Code to work on GPT-5.
  • The company’s terms forbid customers from employing its models to build or train competing AI services, a clause enforced in this latest dispute.
  • OpenAI’s chief communications officer, Hannah Wong, defended the evaluation as standard industry practice and criticized the uneven access between the two firms.
  • Anthropic will sustain limited API availability for benchmarking and safety tests, though the practical impact of these restrictions is still uncertain.
  • OpenAI expects to launch GPT-5 this summer and plans to press ahead with development despite the curtailed API rights.