A spat between the two biggest names in AI assistants spilled into public view this week, and it started with one suspended account.
According to a report carried on MSN, an Anthropic user said their account was suspended after they ran OpenAI's GPT-5.6 Sol on Claude Code, Anthropic's coding tool. The claim spread quickly, and an OpenAI executive identified in the report as Tibo responded with a public dig at Anthropic.
Anthropic's Boris Cherny then stepped in to clarify the circumstances of the ban, per the same report. Shortly afterward, OpenAI reset usage limits for its paid users — a move the report frames as arriving on the heels of the exchange.
The underlying tension is about what you're allowed to run inside someone else's tool. Coding assistants like Claude Code are built to sit in a developer's workflow, and developers naturally want to plug in whichever model works best for a given job. Companies that build those tools, meanwhile, have terms about using them with competing systems. When those two instincts collide, an individual user is the one who gets cut off.
The details available here are thin. The suspension is described as a user's claim, and the substance of Cherny's clarification isn't spelled out in the source. It's also not established that OpenAI's limit reset was a direct response rather than coincidental timing, though the report presents the two as connected.
Why it matters: as AI coding tools become core infrastructure for how software gets written, the rules about which models can run inside which tools stop being fine print and start determining what developers can actually build.