Elevated errors on chat when using Claude Sonnet or Opus 4.5

Anthropic are reporting that a fix has been implemented - we're monitoring error rates to make sure this is fully resolved on our side.

We're seeing higher than usual error rates when generating chat responses if you've selected Claude Sonnet 4.5 or Claude Opus 4.5. This is due to an outage on Claude models. We'll continue to provide updates, and Anthropic are providing updates on their status page: https://status.claude.com/incidents/zt2q0ltqcvbx

You should be able to work around this by switching to another model in Granola chat (e.g. Gemini or GPT).