Aider v0.73.0: o3-mini Support & Reasoning Control

TL;DR

• Full o3-mini support with new --reasoning-effort argument (low/medium/high)
• Better context window handling and improved R1 turn-order enforcement
• Auto-create parent directories for new files; case-insensitive model matching

New

• o3-mini support — Run aider --model o3-mini for OpenAI's faster reasoning model.
• --reasoning-effort argument — Control reasoning depth with low, medium, or high settings.
• Reasoning tag removal — Strip model-specific reasoning tags from responses via remove_reasoning: tagname model setting.
• Auto-create parent directories — New files now create missing parent directories automatically.
• R1 free on OpenRouter — Access DeepSeek R1 free tier with --model openrouter/deepseek/deepseek-r1:free.
• Case-insensitive model matching — Model names now match case-insensitively while preserving original case.

Fixed

• Context window messaging — Improved handling of size limits with better error messages and Ollama-specific guidance.
• R1 turn-order enforcement — Hardened against user/assistant turn order problems that cause R1 errors.
• OpenRouter R1 metadata — Fixed model metadata for openrouter/deepseek/deepseek-r1.

Update with: pip install --upgrade aider-ai

Source: Aider