2026-09-25 21:29 UTC
>Is there any difference between all those coding agents?
what model running in the "agent harness" is going to make the biggest difference. Codex is for OpenAI models, Claude for Anthropic - basically the same from my pov. Then there's stuff like https://pi.dev/ (what I first used and had success with) and https://opencode.ai/
For a while I was (ignorantly) using https://openrouter.ai/ + pi and paying per token API prices (expensive...) before realizing I should just use subscriptions which allow SotA model use for a fraction of the cost
Replies (0)
No replies.