Limit changes

Claude Code updates default model guidance and configuration options

Detected September 23, 2026, 05:15 UTC · Models, usage and limits in Claude Code

Previously, the documentation stated that Sonnet was the fixed default model for all sessions. The updated guidance specifies that the initial model now depends on a user's sign-in method and organizational configurations.

This support article outlines model selection, consumption rates, and operational constraints for developers utilizing Claude Code. In the latest revision, the service removed its blanket statement designating Sonnet as the universal baseline model. Instead, the documentation explains that initial model assignment varies according to how an individual authenticates and the overarching policies established by their organization. Users can also view their presently active engine when running the model selection command, which was not explicitly detailed in the previous text. Furthermore, the documentation expanded its discussion of automatic model fallback to reference Opus 5.5 in addition to Opus 5, and introduced sections addressing quota resets and workspace sharing rules. In practice, programmers can no longer assume every session starts on Sonnet, meaning usage consumption may vary from the outset if an enterprise workspace defaults to a more resource-intensive engine like Opus. However, the documentation does not enumerate the specific defaults linked to particular authentication types, the exact administrative steps required to modify default selections, or the numerical limits and reset schedules governing these accounts.

We compare this page once a day and record every difference. The text above is written by Claude Network from that comparison; none of it is copied from the source.

Source: Models, usage and limits in Claude Code