Scored Runs
Valid agent-driven progress from the current official sweep.
stalled
Gemini 3 Flash via api - Open-ended progress
Best visible milestone: wood 100
4 agency score
28802s
4814 frames
stalled
Gemini 3.1 Pro via api - Open-ended progress
Best visible milestone: wood 25
3 agency score
28802s
3734 frames
complete
Gemini 3.1 Flash Lite via api - Open-ended progress
Best visible milestone: wood 25
3 agency score
21930s
5838 frames
Same Sweep Failure Lanes
Other official models are kept visible for diagnosis, but separated from scored progress.
invalid
Claude Opus 4.7 via api - Open-ended progress
Same sweep failure lane. no agent driven progress, agent stalled
0 agency score
28802s
3326 frames
invalid
GPT-5.5 via api - Open-ended progress
Same sweep failure lane. no agent driven progress, agent stalled
0 agency score
28802s
3080 frames
invalid
Kimi K2.6 via api - Open-ended progress
Same sweep failure lane. no agent driven progress, agent stalled
0 agency score
28802s
1622 frames
invalid
Qwen 3.6 27B via api - Open-ended progress
Same sweep failure lane. no agent driven progress
0 agency score
21935s
6152 frames