Notion restores Anthropic models in Notion AI after disruptions

In short: Notion temporarily turned off Anthropic’s AI models inside Notion AI after errors increased, and it later turned them back on once service recovered.

What happened

Notion, the app many people use for notes and documents, said its Notion AI feature ran into trouble this weekend when it relied on Anthropic’s models. AI models are the “brains” that generate text, similar to an engine that powers a car.

Early Sunday morning, Notion posted a status update saying Anthropic’s Opus 4.7 and 4.8 models had “degraded performance.” In plain terms, they were running poorly and more requests were failing for people who picked those options in Notion AI.

Because of that, Notion said it was disabling all Anthropic models in Notion AI. About 12 hours later, Notion’s head of product, Max Schoening, said the problem was a temporary service disruption and that access had been restored.

Schoening also said he was “astonished” by how much the status post was being reposted on X. He suggested some people wanted the story to be about model quality, when Notion saw it as a short-term outage.

Anthropic also confirmed there was an infrastructure issue, meaning a problem with the underlying computers and systems that run its service (like a power or plumbing issue in a building). Anthropic said the issue caused higher error rates on multiple Claude models for a short period of time and has been resolved.

Why it matters

If you use Notion AI, this is a reminder that features powered by outside services can sometimes go down, even if the main app is still working. For regular users, it can look like the AI “stopped working,” when it is actually a temporary problem with the supplier behind it.

Source: TechCrunch AI