[feature] Add coding model setting — separate model for coding vs orchestration tasks #922
Labels
No labels
area:chat
area:core
area:llm
area:routes
area:tools
bug
documentation
duplicate
enhancement
good first issue
help wanted
invalid
question
refactor
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
sleepy/odysseus#922
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Feature Request
Currently the model settings support:
default_model/default_endpoint_id— main chat modelteacher_model— escalation to more capable modelutility_model— summarization, naming, tidy actionsresearch_model— deep researchtask_model— background tasksMissing: a dedicated coding model. Users who run different models for orchestration (e.g. Qwen for general chat) vs coding (e.g. DeepSeek for code tasks) have no way to configure this separation.
Proposed Changes
New settings keys in
DEFAULT_SETTINGS:coding_model: model ID for coding taskscoding_endpoint_id: endpoint for coding modelcoding_model_fallbacks: fallback chainSettings UI: Add a "Coding" row in the model defaults section (same pattern as utility/teacher/research)
Per-user resolution: Add
coding_endpoint_id,coding_model,coding_model_fallbacksto_PER_USER_KEYSinsrc/settings.pyTool awareness:
chat_with_modeland subagent tools should be able to resolvecodingas a role (same wayteacherresolves to the teacher model)Aliases in
settings_tools.py:"coding model" → "coding_model""coder" → "coding_model"Related