According to Wired magazine, AI company Anthropic has revoked OpenAI's access to its Claude series of AI models, a decision that highlights the increasingly tense relationship between competitors in the artificial intelligence field.
According to insiders, OpenAI had been using the Claude model for internal comparison tests, assessing Claude's performance relative to its own models in multiple dimensions such as programming, writing, and security through specialized tools. This practice made Anthropic dissatisfied, believing that OpenAI violated the terms of service.
Anthropic's spokesperson stated: "OpenAI's technical team was also using our programming tools before the release of GPT-5, which directly violates our terms of service." According to Anthropic's commercial terms, other companies are prohibited from using Claude to build competitive services.
Nevertheless, Anthropic said it would continue to provide OpenAI with access for "benchmark testing and safety evaluation purposes." This retention indicates that even amid intensified business competition, the two companies still maintain some level of cooperation in AI safety assessments.
In response to Anthropic's allegations, OpenAI defended its actions, stating that such practices are "industry standard." OpenAI also expressed disappointment with Anthropic's decision and emphasized, "Our API remains open to them," implying a more open attitude towards collaboration.
This incident is not an isolated one. Anthropic executives have previously shown resistance to providing technical access to competitors. The company's Chief Science Officer, Jared Kaplan, once explained the decision to cut off access to the code editor Windsurf by saying, "It was strange for us to sell Claude to OpenAI." Notably, Windsurf was rumored to be a potential acquisition target for OpenAI, and later was acquired by Cognition company.
This controversy reflects a profound change in the competitive landscape of the AI industry. As large language model technology becomes increasingly mature, companies are striving to protect their technological advantages and prevent competitors from gaining undue advantages. At the same time, it also exposes the complex relationship between AI companies, characterized by both cooperation and competition.
From a business perspective, Anthropic's decision carries significant strategic implications. By limiting competitors' access to Claude, the company attempts to protect its technological moat and prevent OpenAI from leveraging Claude's advantages to improve its models. Such actions can be understood in the context of fierce AI competition but may affect the industry's tradition of openness and collaboration.
For OpenAI, losing direct access to Claude could impact its model development and benchmarking work. Claude's performance on certain tasks is considered to have unique advantages, and losing this reference standard may leave OpenAI lacking important comparative data in the product optimization process.
This incident also raises questions about the collaboration model in the AI industry. In the early stages of rapid technological development, relatively open cooperation between companies benefited the advancement of the entire industry. However, as commercial value becomes more prominent and competition intensifies, this openness is facing challenges.
It is worth noting that despite limited commercial cooperation, the two companies still maintain collaboration in safety assessments. This indicates that AI safety is seen as a common concern beyond commercial competition, reflecting the industry's commitment to responsible AI development.
From a broader industry perspective, this dispute may prompt other AI companies to re-examine their cooperation policies. As technological differentiation becomes increasingly important, more companies may adopt similar protective measures, restricting competitors' access to core technologies.
The dispute between Anthropic and OpenAI signals that the AI industry is transitioning from a relatively open developmental stage to a more closed and competitive mature stage. While this change may reduce technical exchanges within the industry, it could also encourage companies to focus more on independent research and development and innovation, driving the diversification of technology.