The AI Cold War: Anthropic Revokes OpenAI’s Claude API Access Over Terms of Service Dispute
Ddos 
According to a report by WIRED, artificial intelligence company Anthropic has revoked OpenAI’s access to the Claude API, citing violations of its terms of service. The move prohibits OpenAI from continuing to use Claude for debugging its GPT models.
The article states that OpenAI’s technical team is currently preparing the GPT-5 model and had been using the Claude API internally to evaluate the model’s safety and coding capabilities.
Anthropic’s terms of service explicitly prohibit the use of the Claude API to develop competing products. If OpenAI was indeed leveraging Claude in the development of GPT-5, it would constitute a breach of those terms. However, OpenAI denies the allegation.
In its response, OpenAI noted that it is common industry practice to use Claude for benchmarking in order to enhance safety. The company expressed regret over the revocation of access. Anthropic, for its part, stated that it may still allow OpenAI limited access to Claude in the future for standardized benchmarking and security assessments.
It is worth noting that many AI companies routinely evaluate their models against those of their competitors. Such evaluations are vital for improving safety and, more importantly, for comparative performance analysis — a key feature highlighted during the release of new models.
For instance, Anthropic itself has used the OpenAI API to benchmark its Claude series. This makes the decision to ban OpenAI’s access to Claude somewhat perplexing, especially considering that it’s unlikely OpenAI would use Claude API directly to build GPT models.
Furthermore, Anthropic’s recent actions have raised additional suspicions. When OpenAI attempted to acquire WindSurf — an AI programming tool that utilizes APIs from multiple providers, including Claude — Anthropic abruptly terminated WindSurf’s access to Claude’s API without explanation. It remains unclear whether this move was connected to the acquisition effort.
Related Posts:
- OpenAI to Integrate o3 Model into GPT-5, Offering Free Access to All Users
- Claude AI Integrates with Google Workspace
- Smarter Claude AI: Free Users Get Web Search, Voice Mode Coming Soon
- OpenAI Unveils o3 & o4-mini Models, Announces GPT-5 Plans
- OpenAI Delays GPT-5 Launch, Will Release GPT-o3 and o4-mini
Support Our Threat Intelligence
If you find our CVE report and cybersecurity news helpful, consider supporting our work.
Buy Me a Coffee
PayPal
