I've been relying primarily on deepseek-v4-flash for 90% of my work. It sips tokens. that model will run on 128gb. not a cheap configuration for a consumer but within the budget of a developer relying on it for work.
Ive only been using kimi 2.5 and deepseek pro for reviewing PRs for security issues. less than 10% of my workflow requires a full powered frontier model.
I think the issue is overblown by people who think claude code is a good harness and use opus for everything. opencode is objectively better. its much more verbose about what its doing, you have more control when it comes to offloading to subagents with targeted context (crucial for running through larger jobs) and I can swap between codex and open weight models.
However everything in this field is cargo culting. We have absolutely no way to quantify productivity in the real world.
We've had advanced programming languages backed by advanced programming language theory for decades and the most used/ran programming languages in the world are C, PHP and JavaScript, languages held together by duct tape or in the case of C, programing language theory from the 60s.
We have a super minimal JavaScript runtime in the browser to avoid a bloated standard library and then people invent things like leftpad. At the same time basically every major website in the world serves mega bytes of tracking and ad serving libraries.
We all "know" AI makes coders more productive but nobody can do the equivalent of a clinical trial for a major new drug.
Ive only been using kimi 2.5 and deepseek pro for reviewing PRs for security issues. less than 10% of my workflow requires a full powered frontier model.
I think the issue is overblown by people who think claude code is a good harness and use opus for everything. opencode is objectively better. its much more verbose about what its doing, you have more control when it comes to offloading to subagents with targeted context (crucial for running through larger jobs) and I can swap between codex and open weight models.