Story · Cline
Open-sourcing evals for open-weight agents (Cline)
blog post · Story page
Cline says its five most-used models are all open-weight, so it's publishing the evals it uses to judge them on Terminal Bench, tied to model behaviour and token efficiency inside its own harness rather than a bare benchmark score.
In plain words
- Cline published the tests it uses to compare coding artificial intelligence systems whose learned settings are publicly available.
- Cline says its five most-used systems now belong to this publicly available group.
- The tests use Terminal Bench, a set of computer tasks, inside Cline's own coding setup.
- They track system behavior and the amount of text processed, rather than relying only on task completion scores.
- Teams choosing coding systems can now compare performance and text-processing efficiency within the same working setup.
Appeared in
- Aborted agent branches live on in the KV cache, plus StateM's harness scaling
Aug 19, 2026 · in the sections
Subscribe
Get the brief in your inbox
Pick daily, weekly, or both. Nothing is gated either way: every issue is on the site and in the feeds.
- Weekdays at 8:45am IST, one lead story and 6 to 9 items.
- Sundays, an argued synthesis rather than a recap.
- One click to leave, and quiet days say so in the subject line.