Set-aside attempts: attempts cut short by the harness or provider; each was re-run until it completed, and the completed run is the scored record. Most needed one re-run; argon T1 trial 1 needed two, and opus T3 trial 0 and fable T5 trial 0 needed three (one of fable's three is an aborted record kept in the dataset, not shown here). Trials set aside more than once carry -2, -3 ids; each page's footer names its source folder. Not part of the headline.

optobench runs

19 trials · 8 harness failures (no score) · total cost $25.21 · click a trial to step through it

By agent and task (trials shown by the filters)

agent task trialsharness errors submitted all passmean end score costcost / trial
argont1_connectivity3—00 / 30$1.6176$0.5392
argont4_takeover1—00 / 10$0.8973$0.8973
argont5_traces1—11 / 10.999$2.4012$2.4012
argonall tasks5—11 / 50.2$4.9161$0.9832
astrat3_light1—00 / 10$5.0195$5.0195
deepseekt5_traces1—11 / 10.993$0.0857$0.0857
fablet3_light1100 / 0—$1.6442$1.6442
fablet5_traces2111 / 10.998$5.1993$2.5996
fableall tasks3211 / 10.998$6.8435$2.2812
nemotront5_traces2—20 / 20.734$0.2696$0.1348
opust3_light3210 / 10.818$7.3560$2.4520
solt1_connectivity2200 / 0—$0.4537$0.2269
terrat1_connectivity1100 / 0—$0.0917$0.0917
terrat5_traces1100 / 0—$0.1755$0.1755
terraall tasks2200 / 0—$0.2673$0.1336

Each trial has one end grade, computed once after the run ends. Harness failures (the run stopped on a provider or harness error) are listed but have no score: they are left out of "all pass" and the means.

Trials

headlineargont1_connectivityusb_loosetrial 0refused0fail$1.13403471
headlineargont1_connectivityusb_loosetrial 1refused0fail$0.32361523
headlineargont1_connectivityusb_loosetrial 1refused0fail$0.1599815
headlinesolt1_connectivitymask_fliptrial 0harness errorno score—$0.29594564
headlinesolt1_connectivitystale_framestrial 0harness errorno score—$0.15782132
headlineterrat1_connectivityred_led_agedtrial 0harness errorno score—$0.09171934
headlineastrat3_lighttwo_targets_room_lighttrial 0no submission0fail$5.019558103
headlinefablet3_lighttwo_targets_room_lighttrial 1harness errorno score—$1.64422135
headlineopust3_lighttwo_targets_room_lighttrial 0harness errorno score—$3.415254110
headlineopust3_lighttwo_targets_room_lighttrial 0harness errorno score—$1.13933258
headlineopust3_lighttwo_targets_room_lighttrial 0submitted0.818fail$2.801449113
headlineargont4_takeovert_present_faultstrial 0refused0fail$0.89731644
headlineargont5_tracesfrozen_and_off_targettrial 0submitted0.999pass$2.4012168
headlinedeepseekt5_tracesfrozen_and_off_targettrial 0submitted0.993pass$0.08571100
headlinefablet5_tracesfrozen_and_off_targettrial 0harness errorno score—$1.4708014
headlinefablet5_tracesfrozen_and_off_targettrial 0submitted0.998pass$3.7285127
headlinenemotront5_tracesfrozen_and_off_targettrial 0submitted0.838fail$0.1396157
headlinenemotront5_tracesfrozen_and_off_targettrial 1submitted0.631fail$0.1300150
headlineterrat5_tracesfrozen_and_off_targettrial 1harness errorno score—$0.1755024