Providers & models

This is the chapter where OpenCode stops looking like every other terminal agent. Claude Code talks to Anthropic; Codex talks to OpenAI. OpenCode talks to whatever you point it at — Anthropic, OpenAI, Google, Groq, OpenRouter, a model running on your own laptop — and it lets you change that choice without leaving the session. That isn’t a checkbox feature. Once you can put a different model behind every job, you start matching the model to the work, and feedmill is the perfect place to feel why that matters.

feedmill parses RSS, Atom, and a couple of weird JSON-feed variants, and the jobs it throws at you are not the same size. A timezone bug buried in the Atom date parser — where a feed publishing +0530 lands in your reading queue an hour off — is a finicky reasoning problem you want a strong, expensive model chewing on. A sweep across all hundred-odd configured feeds to find which ones still emit a deprecated field is bulk pattern-matching that a cheap, fast model finishes for a fraction of the cost. Run both through the same model and you’re either overpaying for the scan or under-powering the bug. By the end of this chapter you’ll stop doing that.

Here’s the stretch of work, in five moves:

• Wire up more than one provider so the choice exists at all — Wire up more than one provider
• Switch models mid-session with /models, without losing your place — the thread carries over, so the next model inherits the full conversation — Switch models without leaving the session
• Pin a model to each agent so a cheap scout scans while a strong model builds — Give each agent its own model
• Run the same task against two providers and diff the answers to settle a disagreement — Run two providers side by side
• Point OpenCode at a local or custom model for feeds you can’t send to a cloud API — and lean on Zen when you’d rather trust a curated shortlist than benchmark all 75+ providers — Drive it with a local or custom model

Do it as one continuous stretch and the payoff compounds. You’ll wire two providers once, and from then on switching is a keystroke; you’ll set a per-agent model once, and every later chapter’s subagent work inherits the cheap-scout / strong-builder split for free. The single-vendor courses have nothing to teach here because there’s nothing to choose — that’s exactly the lever this chapter hands you. Spend strong reasoning on the timezone bug, cheap throughput on the hundred-feed scan, and keep the private feed on a model that never touches the network, all inside one session. Get the wiring right now and you stop thinking about it; the model just fits the job.