LIBERO-90 task 28 · MolmoAct 2 — LIBERO Evaluation Gallery

Completed task success rate: 42.0%; 21 of 50

Control horizon: 400; Maximum steps

Camera videos: 100; Two per rollout

Evaluation time: 9m 4s; Task process wall time

Canonical catalog

Task definition

Model instruction: close the top drawer of the cabinet
BDDL language: close the top drawer of the cabinet
Problem: libero_kitchen_tabletop_manipulation
Objects: akita_black_bowl_1, ketchup_1, plate_1
Fixtures: kitchen_table, white_cabinet_1
Initial-state predicates: 5
Fixed init states: 50

Exact initial-state predicates

(on akita_black_bowl_1 kitchen_table_akita_black_bowl_init_region)
(on plate_1 kitchen_table_plate_init_region)
(on white_cabinet_1 kitchen_table_white_cabinet_init_region)
(on ketchup_1 kitchen_table_ketchup_init_region)
(open white_cabinet_1_top_region)

Open exact BDDL definition

Simulator truth

Completion criteria

LIBERO checks these BDDL goal predicates after every action. All conditions must be true simultaneously in the same simulator step for the rollout to succeed.

white cabinet 1 top region is closed
(close white_cabinet_1_top_region)

Exact parsed BDDL goal state

(close white_cabinet_1_top_region)

Selected rollout

LIBERO-90 · Task 028 · Episode 1

Failure

Agent view observation.images.image

Eye in hand observation.images.image2

Rollout record

Episode metadata

Link to episode

Episode: 1
Seed: 1000
Init state: 0
Control steps: 400
Wall time: 15s
Outcome: Failure

Evidence browser

All rollouts

Select one episode to mount only its two camera streams in the player.

Episode	Outcome	Seed	Init state	Steps	Video
1	Failure	1000	0	400	172 KB
2	Failure	1001	1	400	172 KB
3	Success	1002	2	233	136 KB
4	Failure	1003	4	400	165 KB
5	Success	1004	5	65	56 KB
6	Failure	1005	7	400	168 KB
7	Failure	1006	8	400	161 KB
8	Failure	1007	9	400	173 KB
9	Failure	1008	10	400	177 KB
10	Failure	1009	11	400	122 KB
11	Success	1010	12	189	125 KB
12	Failure	1011	14	400	179 KB
13	Success	1012	15	48	44 KB
14	Failure	1013	17	400	176 KB
15	Success	1014	18	92	96 KB
16	Failure	1015	20	400	154 KB
17	Failure	1016	21	400	164 KB
18	Success	1017	22	54	47 KB
19	Success	1018	24	249	138 KB
20	Failure	1019	26	400	178 KB
21	Success	1020	27	56	51 KB
22	Success	1021	29	55	48 KB
23	Failure	1022	31	400	163 KB
24	Success	1023	32	113	93 KB
25	Failure	1024	34	400	143 KB
26	Success	1025	35	56	45 KB
27	Failure	1026	37	400	176 KB
28	Success	1027	38	45	42 KB
29	Success	1028	40	147	95 KB
30	Failure	1029	42	400	170 KB
31	Failure	1030	43	400	177 KB
32	Success	1031	44	95	97 KB
33	Success	1032	46	106	109 KB
34	Failure	1033	48	400	135 KB
35	Failure	1034	49	400	153 KB
36	Failure	1035	0	400	166 KB
37	Success	1036	1	74	59 KB
38	Success	1037	3	59	51 KB
39	Success	1038	5	64	54 KB
40	Failure	1039	7	400	164 KB
41	Failure	1040	8	400	136 KB
42	Failure	1041	9	400	149 KB
43	Failure	1042	10	400	154 KB
44	Success	1043	11	99	99 KB
45	Failure	1044	13	400	152 KB
46	Failure	1045	14	400	155 KB
47	Success	1046	15	99	65 KB
48	Success	1047	17	136	110 KB
49	Failure	1048	19	400	146 KB
50	Failure	1049	20	400	174 KB

Integrity

Task attestation

Status: complete
Completed: Aug 16, 2026, 10:00 AM UTC
Manifest SHA-256: d327ede512…71c90d11
Aggregate SHA-256: ddbeea0515…b0dd2d2a
Episode index SHA-256: daa2d29785…dcb0e573
Validation SHA-256: 29414ef386…1a9f4a87
Artifact inventory SHA-256: 0fe2c58af3…7b6c3e42