Artificial Intelligence

Anthropic Blocks OpenAI's Claude API Access

Published AUG 2 2025

Story

Above: Icons of AI chatbot applications. Image credit: Jonathan Raa/NurPhoto/Getty Images

The Facts

Anthropic has revoked OpenAI's API access to its Claude models, citing alleged violations of the terms of service that prohibit using the service to build competing products or reverse-engineer the models.
MINT
Wired
The Information
OpenAI's technical teams allegedly integrated Claude into their internal tools via developer access and systematically compared Claude with their own AI models across coding, creative writing and safety-sensitive content categories.
Wired
THE DECODER
Anthropic spokesperson Christopher Nulty stated that OpenAI's technical staff were using Claude coding tools ahead of GPT-5's launch, which constitutes a direct violation of the commercial terms of service. OpenAI claims its use of the tools met "industry standard[s]."
MINT
Wired
The Information
Anthropic announced that it will continue to allow API access for benchmarking and safety evaluation purposes, describing this as standard industry practice, although it did not clarify OpenAI's future eligibility.
MINT
Wired
Anthropic previously restricted AI coding startup Windsurf's access to Claude models in June, following reports that OpenAI planned to acquire the company for approximately $3 billion.
Forbes
TechCrunch
X

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Narrative A

Anthropic's decision to cut OpenAI's access is a necessary measure to protect intellectual property and enforce contracts. OpenAI violated its terms by using Claude to develop competing products rather than for legitimate benchmarking purposes. The timing preceding GPT-5's release makes this behavior particularly suspicious and warrants immediate attention.

PCMAG

Narrative B

This move represents anticompetitive behavior that stifles industry collaboration and innovation. Benchmarking competitor models is a standard practice for safety and progress; yet, Anthropic maintains access to OpenAI's systems while blocking reciprocal access. Such tactics harm the broader AI development ecosystem.

Opentools.ai