Anthropic Terminates OpenAI's Access to Claude Models Amid Safety Concerns

Anthropic has revoked OpenAI's access to its Claude AI models due to a violation of Anthropic's terms of service. OpenAI was using Claude to compare its performance to its own models, which Anthropic deemed a direct violation. Anthropic's CEO Jared Kaplan previously justified cutting off access to Windsurf, an OpenAI acquisition target, by saying it would be odd to sell Claude to a competitor.

Anthropic has revoked OpenAI's access to its Claude AI models due to a violation of Anthropic's terms of service. The move comes as OpenAI prepares to launch GPT-5, a new AI model expected to enhance coding capabilities. The incident highlights the growing tensions and competitive practices in the AI industry.

According to sources familiar with the matter, OpenAI was using Claude to compare its performance to its own models, a practice that Anthropic deemed a direct violation of its commercial terms of service. These terms prohibit customers from using the service to build competing products or services, including to train competing AI models, or to reverse engineer or duplicate the services [1].

OpenAI's chief communications officer, Hannah Wong, defended the company's actions, stating that it is industry standard to evaluate other AI systems to benchmark progress and improve safety. However, she expressed disappointment at Anthropic's decision to cut off their API access [1].

The timing of this incident is particularly noteworthy as OpenAI is reportedly preparing to release GPT-5 next week. The new model is expected to launch with "auto" and reasoning modes, potentially making it more competitive in the AI coding industry [2].

This is not an isolated incident in the tech industry. Similar actions have been taken by other companies, such as Facebook revoking Twitter-owned Vine's API access and Salesforce restricting competitors' access to certain data through the Slack API. Anthropic itself had previously restricted AI coding startup Windsurf's access to its models when rumors of an OpenAI acquisition surfaced [1].

Anthropic's CEO Jared Kaplan previously justified cutting off access to Windsurf by saying it would be odd to sell Claude to a competitor. This move underscores the competitive nature of the AI industry and the challenges companies face in balancing collaboration and the protection of proprietary technologies [3].

References:
[1] https://theoutpost.ai/news-story/anthropic-revokes-open-ai-s-access-to-claude-api-amid-gpt-5-launch-preparations-18573/
[2] https://techcrunch.com/2025/08/02/anthropic-cuts-off-openais-access-to-its-claude-models/
[3] https://www.wired.com/story/anthropic-revokes-openais-access-to-claude/