Cline added stealth model Ox Alpha as a free option. Early benches show a slight edge over Fable and GPT, and it already drove 8% of all Cline inference within 48 hours. On a real Cline-repo bug, both models fixed the issue, but Ox used roughly 3x fewer output tokens by acting on its first conclusion instead of re-deriving it.

Key Takeaways

  • โœ“Install with npm i -g cline and pick Ox Alpha under Free in /models; also on VS Code and JetBrains
  • โœ“Fastest-growing model in Cline history, already 8% of all inference after 48 hours
  • โœ“On a real repo bug both models shipped a fix, but Ox used ~3x fewer output tokens by skipping re-verification loops