OpenCode makes Gemini 3.1 Pro usable

This is really a story about reliability, not novelty: the winning AI coding tool is often the one that can actually route the model you want without friction. OpenCode looks stronger here because its model-agnostic design is built around swapping providers instead of locking developers into one UX.

• –OpenCode positions itself as an open-source AI coding agent for terminal, desktop, and IDE use, which gives developers one workflow across multiple surfaces
• –Its docs and homepage emphasize support for 75+ LLM providers, plus direct access to Claude, GPT, Gemini, local models, GitHub Copilot, and ChatGPT accounts
• –The Zen docs explicitly argue that model quality depends on provider setup, which matches the video's claim that Gemini 3.1 Pro becomes workable when routed differently
• –With more than 100K GitHub stars and hundreds of contributors, OpenCode is starting to look like a credible open-source counterweight to closed AI coding tools