Download

0.0.121

April 26, 2025

• Attempt to fix renderer crash issue when conversation history and messages are very long by fallback to a simpler markdown renderer for long messages

Thanks to techno. for the feedback and suggestions!

0.0.120

April 17, 2025

• Added support for OpenAI o3 and o4-mini models
  • They are both quite good at following instructions
  • o3 is very expensive ($40/million tokens output), o4-mini is much cheaper ($4.4/million tokens output)
  • They can occasional output diff/patch format output, you can suppress it by adding Do NOT output diff/patch format. to your custom instructions
  • Note that you might need to verify your organization to access these models

0.0.119

April 15, 2025

• Added support for GPT-4.1 models (GPT-4.1, GPT-4.1-mini, GPT-4.1-nano)
  • GPT-4.1 is much better at following instructions than existing models such as Claude 3.7 Sonnet and Gemini 2.5 Pro
  • GPT-4.1 is good at giving diff format output, producing only changed lines
  • GPT-4.1 is also much faster and cheaper than existing models
  • We might switch the default model from Claude 3.5 Sonnet to GPT-4.1 in the near future, pending more testing