What
When you switch models with /model, Claude Code first checks the model with Anthropic. That check can be refused with a 429, the error meaning you have hit a usage limit.
- When the 429 is not about needing credits and the limit is the
five_hourorseven_dayone, the result is now markedsharedLimitReached: true. - A switch that would otherwise fail with
invalid_modelis now allowed whensharedLimitReachedis true and the caller passesacceptSharedLimitRefusal. This works like the existing path for credits-required refusals. - The model check now passes
acceptSharedLimitRefusal: true, alongside the existingacceptSessionModeRefusal.
Before this release, such a switch was rejected as invalid_model unless credits or another condition allowed it.
Why
Hitting your plan's usage limit during the check no longer makes a valid model look invalid. The switch goes ahead, and the limit can be reported for what it is.
Names in the bundle/model
The entry above is what we published on the day. These lines were added later, as Anthropic's own pages caught up, and they sit beside the original rather than replacing it.
Confirmed since
Anthropic's documentation has since written up /model, on Choose a permission mode.
| Has [`ANTHROPIC_DEFAULT_SONNET_MODEL`](/docs/en/model-config#environment-variables) set | Its value, or Sonnet 5 if its value is Claude Sonnet 5.5 or Claude Opus 5.5 |permission-modes see the edit
Added since
A small documentation edit on Agent SDK reference - Python touched a line naming /model after this was published.
| `set_model(model)` | Change the model for the current session. Pass `None` to reset to [Claude Code's default model](/docs/en/model-config). A session in `auto` mode [changes to `default` mode](/docs/en/agent-sdk/permissions#auto-mode-au…agent-sdk/python see the edit
One source agreesOne thing we can check says the same as this entry.
Anthropic's documentation agrees
A small documentation edit on Agent SDK reference - Python touched a line naming /model after this was published.