Notion briefly cut off its users from Anthropic's AI models on Sunday after the productivity software company reported a spike in failures tied to Claude Opus 4.7 and 4.8. In a post on X early Sunday morning, Notion said the two models were "experiencing degraded performance" and were returning errors at higher-than-normal rates when selected inside Notion AI. The company responded by disabling access to all Anthropic models across the platform, a move that touched a wide swath of its user base.
The post quickly picked up traction, generating roughly 1,200 reposts within hours — a reaction that drew a pointed response from Notion's head of product, Max Schoening. Roughly twelve hours after the initial notice, Schoening said he was "astonished" by the volume of users reposting the update, suggesting many were hoping to spin the incident into a story about declining model quality. He insisted the disruption was purely a temporary infrastructure issue, comparing it to outages that routinely hit Notion itself, GitHub, AWS, and other major services. By the time he posted, Notion had already restored access to Anthropic's models.
Anthropic echoed that explanation through a spokesperson, who confirmed that a "brief infrastructure issue" had caused elevated error rates across multiple Claude models for a short window before being resolved. The company thanked users for their patience during the fix. The incident ended without any lasting impact on service, but it offered a small window into how dependent third-party AI products have become on a handful of model providers — and how quickly users notice when those connections falter.