E5 · Scenario sweep: 65 test scenarios × 1 trial on the frozen code (no change to what the model sees); terra and sol, experiments/sweep1_openai. All-pass: Terra 44/65, Sol 48/65. Not part of the headline.

optobench runs

130 trials · total cost $27.97 · 2 Sol T3 episodes ended on an OpenAI rate limit (shown as harness failures; fails in the all-pass counts) · not on this page: 100 attempts that ended on an empty OpenAI account ($0.23, re-run until completed); they are in the dataset's experiments/sweep1_openai/setaside/ · 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
solt1_connectivity14—149 / 140.84$4.4179$0.3156
solt2_setup15—159 / 150.964$1.6137$0.1076
solt3_light1621411 / 160.923$6.2606$0.3913
solt4_takeover11—1110 / 110.909$3.2095$0.2918
solt5_traces9—99 / 90.998$2.4967$0.2774
solall tasks6526348 / 650.923$17.9984$0.2769
terrat1_connectivity14—1410 / 140.984$2.2905$0.1636
terrat2_setup15—158 / 150.951$1.0392$0.0693
terrat3_light16—1612 / 160.935$3.0514$0.1907
terrat4_takeover11—116 / 110.905$2.0538$0.1867
terrat5_traces9—98 / 90.98$1.5364$0.1707
terraall tasks65—6544 / 650.951$9.9713$0.1534

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

headlinesolt1_connectivitydead_culturetrial 0submitted1pass$0.39694571
headlinesolt1_connectivityhdmi_loosetrial 0submitted1pass$0.25172745
headlinesolt1_connectivityled_power_loosetrial 0submitted0.846fail$0.24202844
headlinesolt1_connectivityloose_wiretrial 0submitted1pass$0.38253766
headlinesolt1_connectivitylow_disktrial 0submitted0fail$0.16181827
headlinesolt1_connectivitymask_fliptrial 0submitted1pass$0.38784668
headlinesolt1_connectivitynonetrial 0submitted1pass$0.33823658
headlinesolt1_connectivityoperator_awaytrial 0submitted1pass$0.29203550
headlinesolt1_connectivityprojector_standbytrial 0submitted1pass$0.32063055
headlinesolt1_connectivityred_led_agedtrial 0submitted0.957fail$0.40633759
headlinesolt1_connectivityroom_lighttrial 0submitted0fail$0.36124865
headlinesolt1_connectivitystale_framestrial 0submitted1pass$0.27563049
headlinesolt1_connectivitytrigger_cabletrial 0submitted0.96fail$0.30683352
headlinesolt1_connectivityusb_loosetrial 0submitted1pass$0.29453048
headlineterrat1_connectivitydead_culturetrial 0submitted0.883fail$0.21023661
headlineterrat1_connectivityhdmi_loosetrial 0submitted1pass$0.17001934
headlineterrat1_connectivityled_power_loosetrial 0submitted1pass$0.12312742
headlineterrat1_connectivityloose_wiretrial 0submitted1pass$0.19303253
headlineterrat1_connectivitylow_disktrial 0submitted1pass$0.13132439
headlineterrat1_connectivitymask_fliptrial 0submitted0.969fail$0.17173855
headlineterrat1_connectivitynonetrial 0submitted1pass$0.12621939
headlineterrat1_connectivityoperator_awaytrial 0submitted1pass$0.12762137
headlineterrat1_connectivityprojector_standbytrial 0submitted1pass$0.13402542
headlineterrat1_connectivityred_led_agedtrial 0submitted0.969fail$0.17062448
headlineterrat1_connectivityroom_lighttrial 0submitted1pass$0.22413352
headlineterrat1_connectivitystale_framestrial 0submitted1pass$0.22153862
headlineterrat1_connectivitytrigger_cabletrial 0submitted0.96fail$0.13292643
headlineterrat1_connectivityusb_loosetrial 0submitted1pass$0.15432545
headlinesolt2_setupcalibration_missingtrial 0submitted0.956fail$0.08931522
headlinesolt2_setupcalibration_staletrial 0submitted0.867fail$0.09241622
headlinesolt2_setupdense_culturetrial 0submitted1pass$0.09731523
headlinesolt2_setupdistortion_densetrial 0submitted0.867fail$0.09211420
headlinesolt2_setupfew_markerstrial 0submitted1pass$0.21912433
headlinesolt2_setuphdmi_loosetrial 0submitted1pass$0.07071218
headlinesolt2_setuplens_distortiontrial 0submitted1pass$0.11111725
headlinesolt2_setuploose_wiretrial 0submitted1pass$0.10102027
headlinesolt2_setupmask_fliptrial 0submitted0.956fail$0.08991320
headlinesolt2_setupmissing_uneventrial 0submitted0.956fail$0.15201929
headlinesolt2_setupnonetrial 0submitted1pass$0.07751317
headlinesolt2_setupoperator_awaytrial 0submitted1pass$0.09181825
headlinesolt2_setupprojector_standbytrial 0submitted1pass$0.11941928
headlinesolt2_setuproom_lighttrial 0submitted0.867fail$0.08791520
headlinesolt2_setupuneven_illuminationtrial 0submitted1pass$0.12241425
headlineterrat2_setupcalibration_missingtrial 0submitted0.956fail$0.04511217
headlineterrat2_setupcalibration_staletrial 0submitted1pass$0.06661523
headlineterrat2_setupdense_culturetrial 0submitted1pass$0.08521725
headlineterrat2_setupdistortion_densetrial 0submitted0.867fail$0.06971625
headlineterrat2_setupfew_markerstrial 0submitted0.956fail$0.12632536
headlineterrat2_setuphdmi_loosetrial 0submitted1pass$0.04891420
headlineterrat2_setuplens_distortiontrial 0submitted0.867fail$0.05381421
headlineterrat2_setuploose_wiretrial 0submitted1pass$0.06041624
headlineterrat2_setupmask_fliptrial 0submitted1pass$0.06251420
headlineterrat2_setupmissing_uneventrial 0submitted0.948fail$0.11321932
headlineterrat2_setupnonetrial 0submitted1pass$0.05661520
headlineterrat2_setupoperator_awaytrial 0submitted1pass$0.05091521
headlineterrat2_setupprojector_standbytrial 0submitted1pass$0.06131924
headlineterrat2_setuproom_lighttrial 0submitted0.867fail$0.03981519
headlineterrat2_setupuneven_illuminationtrial 0submitted0.811fail$0.09891929
headlinesolt3_lightclose_neighbourstrial 0submitted0.977pass$0.30303768
headlinesolt3_lightcluster_fast_bleachtrial 0submitted0.986pass$0.30953759
headlinesolt3_lightconvergingtrial 0harness errorno score—$0.747154101
headlinesolt3_lightconverging_led_lowtrial 0submitted0.973pass$0.54795993
headlinesolt3_lightcrosstalk_two_targetstrial 0submitted0.828fail$0.27753656
headlinesolt3_lightdead_culturetrial 0submitted0.976pass$0.36305078
headlinesolt3_lightfast_bleachtrial 0submitted0.667fail$0.30514264
headlinesolt3_lightled_lowtrial 0submitted0.981pass$0.34773969
headlinesolt3_lightloose_wiretrial 0submitted0.968pass$0.33803866
headlinesolt3_lightnonetrial 0submitted1pass$0.24883252
headlinesolt3_lightred_led_agedtrial 0submitted0.986pass$0.37343971
headlinesolt3_lightroom_lighttrial 0submitted0.667fail$0.38723964
headlinesolt3_lightstale_framestrial 0harness errorno score—$0.48884576
headlinesolt3_lightstrong_crosstalktrial 0submitted0.992pass$0.27063456
headlinesolt3_lightstrong_inputtrial 0submitted0.94pass$0.30443964
headlinesolt3_lighttwo_targets_room_lighttrial 0submitted0.98pass$0.648760103
headlineterrat3_lightclose_neighbourstrial 0submitted0.998pass$0.12692949
headlineterrat3_lightcluster_fast_bleachtrial 0submitted0.998pass$0.16303156
headlineterrat3_lightconvergingtrial 0submitted0.943pass$0.358185127
headlineterrat3_lightconverging_led_lowtrial 0submitted0.667fail$0.375576112
headlineterrat3_lightcrosstalk_two_targetstrial 0submitted0.72fail$0.17123255
headlineterrat3_lightdead_culturetrial 0submitted0.975pass$0.14103859
headlineterrat3_lightfast_bleachtrial 0submitted0.991pass$0.13062950
headlineterrat3_lightled_lowtrial 0submitted0.898fail$0.16193157
headlineterrat3_lightloose_wiretrial 0submitted0.984pass$0.14513450
headlineterrat3_lightnonetrial 0submitted0.981pass$0.12942746
headlineterrat3_lightred_led_agedtrial 0submitted0.959pass$0.17473563
headlineterrat3_lightroom_lighttrial 0submitted0.967fail$0.18273359
headlineterrat3_lightstale_framestrial 0submitted0.969pass$0.20203565
headlineterrat3_lightstrong_crosstalktrial 0submitted0.976pass$0.18532855
headlineterrat3_lightstrong_inputtrial 0submitted0.953pass$0.14733154
headlineterrat3_lighttwo_targets_room_lighttrial 0submitted0.977pass$0.25674176
headlinesolt4_takeovercleantrial 0submitted1pass$0.27032240
headlinesolt4_takeoverdead_culture_swaptrial 0submitted1pass$0.31582748
headlinesolt4_takeoverdisk_note_finetrial 0submitted0fail$0.23042440
headlinesolt4_takeoverdisk_triagetrial 0submitted1pass$0.24372337
headlinesolt4_takeoverflip_grid_checktrial 0submitted1pass$0.37293963
headlinesolt4_takeoverfrozen_counterstrial 0submitted1pass$0.27782542
headlinesolt4_takeoverloose_jumpertrial 0submitted1pass$0.21241429
headlinesolt4_takeovermoved_sampletrial 0submitted1pass$0.32452346
headlinesolt4_takeovert_dead_jumpertrial 0submitted1pass$0.26391939
headlinesolt4_takeovert_moved_frozentrial 0submitted1pass$0.31212646
headlinesolt4_takeovert_present_faultstrial 0submitted1pass$0.38563662
headlineterrat4_takeovercleantrial 0submitted1pass$0.13341427
headlineterrat4_takeoverdead_culture_swaptrial 0submitted1pass$0.32954076
headlineterrat4_takeoverdisk_note_finetrial 0submitted1pass$0.12491732
headlineterrat4_takeoverdisk_triagetrial 0submitted0.998fail$0.16622239
headlineterrat4_takeoverflip_grid_checktrial 0submitted1pass$0.18063150
headlineterrat4_takeoverfrozen_counterstrial 0submitted1pass$0.14662137
headlineterrat4_takeoverloose_jumpertrial 0submitted0.971fail$0.13651631
headlineterrat4_takeovermoved_sampletrial 0submitted1pass$0.17141334
headlineterrat4_takeovert_dead_jumpertrial 0submitted0.988fail$0.15511738
headlineterrat4_takeovert_moved_frozentrial 0submitted0.996fail$0.28542354
headlineterrat4_takeovert_present_faultstrial 0submitted0fail$0.22402747
headlinesolt5_tracesbleachedtrial 0submitted0.999pass$0.3213131
headlinesolt5_tracesbleached_and_droppedtrial 0submitted0.998pass$0.3091125
headlinesolt5_tracescleantrial 0submitted0.998pass$0.2602124
headlinesolt5_tracesdropped_and_lighttrial 0submitted0.997pass$0.2590122
headlinesolt5_traceseverythingtrial 0submitted0.997pass$0.3250126
headlinesolt5_tracesfrozen_and_off_targettrial 0submitted0.999pass$0.2714122
headlinesolt5_tracesfrozen_stretchtrial 0submitted1pass$0.2783126
headlinesolt5_traceslight_and_frozentrial 0submitted0.999pass$0.2715124
headlinesolt5_tracesoff_targettrial 0submitted0.998pass$0.2010117
headlineterrat5_tracesbleachedtrial 0submitted0.847fail$0.1739125
headlineterrat5_tracesbleached_and_droppedtrial 0submitted0.993pass$0.2199133
headlineterrat5_tracescleantrial 0submitted0.999pass$0.1496120
headlineterrat5_tracesdropped_and_lighttrial 0submitted0.995pass$0.1782125
headlineterrat5_traceseverythingtrial 0submitted0.998pass$0.2003127
headlineterrat5_tracesfrozen_and_off_targettrial 0submitted0.997pass$0.1803126
headlineterrat5_tracesfrozen_stretchtrial 0submitted0.999pass$0.1813132
headlineterrat5_traceslight_and_frozentrial 0submitted0.999pass$0.1242116
headlineterrat5_tracesoff_targettrial 0submitted0.997pass$0.1286125