Aider v0.64.0: /editor command, gpt-4o-2024-11-20 support, o1 streaming by default
Adds a /editor slash command that opens your system editor to compose a prompt, full support for the gpt-4o-2024-11-20 model, and turns on streaming for o1 models by default. Also adds global extra_params for litellm and several git/file-handling fixes.
PUBLISHED2024-11-21
OBSERVED2026-08-11
AGE1y
SOURCES1
- New
/editorcommand opens the system editor to write a prompt - Full support for
gpt-4o-2024-11-20 - o1 models now stream by default
/runand suggested shell commands confirm how many lines of output were added to the chat- Optional multiline input tags with matching closing tags
- Model settings gain a global
extra_paramspassed tolitellm.completion() - Architect mode now offers to add files the LLM suggests
- New
--show-release-notescontrols release-note display on first run of a new version --readcan now load directories recursively- 1% of users prompted to opt in to analytics
- Fixes: fuzzy model-name matching (including when litellm provider info is missing), git-ignored file handling, git error handling, Timeout exception for provider timeouts, exponential backoff when editor file locks block writes, empty dict cached on model-metadata download failure to delay retry
- Updated Qwen 2.5 Coder 32B model configuration and dependency versions for a litellm ollama fix
COMMUNITY
No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.