Particle.news
Download on the App Store
9 ARTICLES
·
last wk.

Anthropic Bars OpenAI From Full Claude API Access

Anthropic said its action followed use of the coding tool to benchmark features of the next-generation AI model, with only limited endpoints reserved for safety and performance testing.

Image
anthropic logo on orange background
Image
Anthropic

Overview

  • On August 1–2, Anthropic revoked OpenAI’s general API access after finding its engineers used Claude Code in breach of its commercial terms.
  • Anthropic’s terms explicitly bar customers from using Claude to train or reverse-engineer competing AI models.
  • The company will continue to provide restricted API pathways for OpenAI to conduct industry-standard safety evaluations and benchmarking.
  • OpenAI described the cutoff as disappointing and noted that Anthropic still has full access to OpenAI’s API.
  • The move underscores intensifying competition over model capabilities as OpenAI prepares to launch GPT-5 and mirrors past tech rivals’ use of API restrictions.