Add feature flag to enable personalised context window for AI models
What does this MR do and why?
This MR introduces a feature flag ai_per_model_context_window that enables configurable context window sizes per AI model, allowing models like Claude Sonnet 4.5 (1M tokens) and GPT-5 (400K tokens) to use their respective limits instead of the current hardcoded 400K default.
References
- Parent issue: gitlab-org/modelops/applied-ml/code-suggestions/ai-assist#1515
- MR requiring the feature flag: gitlab-org/modelops/applied-ml/code-suggestions/ai-assist!3810
MR acceptance checklist
Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.