68.1%pass rate, 95% interval 60.0–76.4
wozcode 2.1.86, Lean contextwozcode 2.1.86, Cache usewozcode 2.1.86, Tests before finishingwozcode 2.1.86, No loopswozcode 2.1.86, Speed
wozcode
wozcodewozcode's pass rate on every board it ran on. Scores on different boards don't add up, so this page doesn't add them.
On each board
Terminal-Bench 2.0
terminal-bench@2.0Claude Opus 4.62.1.86, full, 89 tasks, 5 attempts, open
Ranks 3–7
8 harnesses68.1%60.0–76.4wozcode minus Terminus 2 2.0.0: +5.2 points, 95% interval −1.3 to +12.1, over 89 tasks both ran; too close to call.
8 harnesses68.1%60.0–76.4wozcode minus Terminus 2 2.0.0: +5.2 points, 95% interval −1.3 to +12.1, over 89 tasks both ran; too close to call.
How wozcode works
Each hexagon shows how wozcode works, ranked among the harnesses on that board: outward is better.
Compare with another harness
wozcode 2.1.86, Lean contextwozcode 2.1.86, Cache usewozcode 2.1.86, Tests before finishingwozcode 2.1.86, No loopswozcode 2.1.86, Speedmeta-harness 1.1.0, Lean contextmeta-harness 1.1.0, Cache usemeta-harness 1.1.0, Tests before finishingmeta-harness 1.1.0, Direct pathmeta-harness 1.1.0, No loopsmeta-harness 1.1.0, Speed
- wozcode 2.1.86 · 68.1%
- meta-harness 1.1.0 · 76.4%