Bug Report
Description
In the model picker/dropdown, Claude models are displayed with a context size label of 200k (long context), but the actual context window used when sending requests is shorter than expected. All Claude models show the same 200k figure regardless of their actual supported context sizes.
Steps to Reproduce
- Open the Copilot Chat panel
- Click the model picker dropdown
- Observe the context size shown for Claude models (e.g. claude-sonnet-4.6, claude-opus-4.8)
Expected Behavior
Each model should display its correct context size. For example:
claude-sonnet-4.6 → actual context window (not 200k)
claude-opus-4.8 → actual context window (not 200k)
The context size label should accurately reflect the context window actually available/used for that model.
Actual Behavior
All Claude models (including \claude-sonnet-4.6\ and \claude-opus-4.8) are labeled with 200k context in the dropdown, even if they do not support a 200k context window. This can mislead users into thinking they can use more context than is actually available.
Impact
Users may attempt to send large contexts assuming the model supports 200k tokens, only to encounter errors or truncated responses.
Environment
- Plugin: GitHub Copilot for Eclipse
- Component: Model picker UI / model metadata
Screenshots

Bug Report
Description
In the model picker/dropdown, Claude models are displayed with a context size label of 200k (long context), but the actual context window used when sending requests is shorter than expected. All Claude models show the same 200k figure regardless of their actual supported context sizes.
Steps to Reproduce
Expected Behavior
Each model should display its correct context size. For example:
claude-sonnet-4.6→ actual context window (not 200k)claude-opus-4.8→ actual context window (not 200k)The context size label should accurately reflect the context window actually available/used for that model.
Actual Behavior
All Claude models (including \claude-sonnet-4.6\ and \claude-opus-4.8) are labeled with 200k context in the dropdown, even if they do not support a 200k context window. This can mislead users into thinking they can use more context than is actually available.
Impact
Users may attempt to send large contexts assuming the model supports 200k tokens, only to encounter errors or truncated responses.
Environment
Screenshots