You'll noticeTier: how much it should matter to you
3Useful: my rating, 1 to 5
4Signal: worth watching, 1 to 5
ModelsArea: what it touches
ImprovementsKind: in v2.1.287,
What probably matters to youSection of the release
What
Claude Code keeps a built-in list of models with details about each one. Two Sonnet entries in that list have changed: the one that falls back to claude-sonnet-4-6, and the one that falls back to claude-sonnet-5 on third-party providers. Third-party providers are the cloud services that can host Claude models: Amazon Bedrock, Google Vertex AI and Microsoft Foundry.
Both entries used to say that Bedrock, Vertex and Foundry natively support a context window of 1 million tokens. That setting has been removed. The context window is how much text the model can consider at once. Both entries still list a 1-million-token context window, native 1M support and support for the 1M beta.
Why
If you use these Sonnet models through Bedrock, Vertex or Foundry, Claude Code may now handle their large context window differently.
How sure we are
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubtIt is not clear how Claude Code treats a model with no native 1M setting for third-party providers.
