OpenAI has decided to end its contract with the AI coding tool Cursor, effective November 12, 2026. The decision follows SpaceX's acquisition of Cursor, which OpenAI claims has raised concerns about the company's commitment to honoring contractual terms. According to OpenAI, the move is a direct response to Elon Musk's companies having a track record of breaking contracts, including past instances involving Twitter and other AI models. 'It boils down to trust,' said Thibault Sottiaux, OpenAI's representative, emphasizing the company's concerns about the reliability of SpaceX's adherence to service agreements. OpenAI will still allow users to access GPT models through Cursor using their own API keys and continue providing access via its IDE extensions for Cursor. Sottiaux clarified that this action is specifically targeted at Musk and not the broader AI coding tool landscape or its developer community.
Anthropic, a competitor in the AI space, has positioned itself as a more reliable partner, citing its long-standing relationship with Cursor since the Sonnet 3.5 release. Tom Brown, Anthropic's co-founder and Chief Compute Officer, stated that the company will continue to expand compute capacity for its Claude models through Cursor. However, this stance has been seen as ironic, given Anthropic's own history of revoking API access from partners, including blocking Windsurf after its planned sale to OpenAI and restricting OpenAI's access for unauthorized use of Claude in benchmarks. OpenAI's decision to cut off Cursor is part of a larger dispute with Musk, which has also involved legal battles over data usage and contractual breaches. The AI lab has accused Musk of violating terms of service by using data from other labs' models to train Grok, a practice known as distillation. OpenAI is also framing this as a safety issue, particularly with the upcoming Astra model, which is said to have advanced cyber capabilities, requiring a new level of accountability from partners.
Cursor co-founder Michael Truell responded to the news, stating that OpenAI models account for only about five percent of Cursor's AI traffic. He emphasized the company's long-standing partnership with OpenAI and its trust in the platform as neutral infrastructure. Truell noted that Cursor is currently in discussions with OpenAI's team to find a resolution. The situation highlights the ongoing tensions in the AI industry, where trust and contractual compliance are central issues, especially with major players like OpenAI and Anthropic navigating complex relationships and legal challenges. OpenAI's recent actions also reflect its broader concerns about the potential risks of AI, which have been a consistent focus of the company's public statements and policy decisions.
Source: thedecoder
