Dominik Biedebach @bdbch.com · May 16

I'm currently test-running Opencode Go (10$ sub) and I'm honestly amazed. It does lack models like OpenAI's or Anthropics but the low-cost models really amazed me for how much they offer you. This is my usage after 2 weeks of heavy usage.

2 likes 1 replies

?

Replies

Dominik Biedebach · May 16

I mostly run more complex planning tasks through the more expensive models like GLM or Kimi K2 and then pass the implementation plan over to something like Deepseek V4 Pro or Minimax and for most usecases I just fallback to Deepseek V4 Fast which costs like 0.005$ per tool run.