MolmoAct 2 Evaluation gallery
IN PROGRESS · PROVISIONAL Only fully completed, validated task shards are shown. Campaign coverage and aggregate rates can change.
Paper protocol Task 003 Validated task

put the black bowl in the bottom drawer of the cabinet and close it

Completed task success rate
94.0%
47 of 50
Control horizon
520
Maximum steps
Camera videos
100
Two per rollout
Evaluation time
8m 41s
Task process wall time

Canonical catalog

Task definition

Model instruction
put the black bowl in the bottom drawer of the cabinet and close it
BDDL language
put the black bowl in the bottom drawer of the cabinet and close it
Problem
libero_kitchen_tabletop_manipulation
Objects
akita_black_bowl_1, wine_bottle_1
Fixtures
kitchen_table, white_cabinet_1, wine_rack_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 wine_bottle_1 kitchen_table_wine_bottle_init_region)
(on white_cabinet_1 kitchen_table_white_cabinet_init_region)
(on wine_rack_1 kitchen_table_wine_rack_init_region)
(open white_cabinet_1_bottom_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.

  1. white cabinet 1 bottom region is closed

    (close white_cabinet_1_bottom_region)

  2. akita black bowl 1 is in white cabinet 1 bottom region

    (in akita_black_bowl_1 white_cabinet_1_bottom_region)
Exact parsed BDDL goal state 
(close white_cabinet_1_bottom_region)
(in akita_black_bowl_1 white_cabinet_1_bottom_region)

Selected rollout

LIBERO-10 · Task 003 · Episode 1

Success
Agent view observation.images.image
Eye in hand observation.images.image2
Videos are not loaded yet.

Rollout record

Episode metadata

Link to episode
Episode
1
Seed
1000
Init state
0
Control steps
213
Wall time
10s
Outcome
Success

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 Success 1000 0 213 198 KB
2 Success 1001 2 232 198 KB
3 Success 1002 4 219 200 KB
4 Success 1003 6 219 200 KB
5 Success 1004 8 256 208 KB
6 Success 1005 10 215 198 KB
7 Success 1006 12 215 197 KB
8 Success 1007 14 212 188 KB
9 Success 1008 16 214 186 KB
10 Success 1009 18 248 211 KB
11 Success 1010 20 239 206 KB
12 Success 1011 22 215 199 KB
13 Success 1012 24 223 206 KB
14 Success 1013 26 224 209 KB
15 Success 1014 28 216 196 KB
16 Success 1015 30 218 197 KB
17 Success 1016 32 223 198 KB
18 Success 1017 34 221 197 KB
19 Success 1018 36 214 198 KB
20 Success 1019 38 256 231 KB
21 Success 1020 40 347 227 KB
22 Success 1021 42 222 201 KB
23 Success 1022 44 203 191 KB
24 Success 1023 46 247 215 KB
25 Success 1024 48 233 199 KB
26 Success 1025 0 212 205 KB
27 Success 1026 2 252 218 KB
28 Success 1027 4 224 193 KB
29 Success 1028 6 202 187 KB
30 Success 1029 8 227 196 KB
31 Failure 1030 10 520 264 KB
32 Success 1031 11 223 195 KB
33 Success 1032 13 427 238 KB
34 Success 1033 15 214 199 KB
35 Success 1034 17 196 182 KB
36 Success 1035 19 217 189 KB
37 Success 1036 21 253 232 KB
38 Failure 1037 23 520 231 KB
39 Success 1038 24 248 219 KB
40 Success 1039 26 211 202 KB
41 Success 1040 28 223 194 KB
42 Success 1041 30 242 205 KB
43 Success 1042 32 217 195 KB
44 Success 1043 34 244 214 KB
45 Success 1044 36 221 194 KB
46 Success 1045 38 237 212 KB
47 Success 1046 40 205 192 KB
48 Success 1047 42 258 207 KB
49 Success 1048 44 201 189 KB
50 Failure 1049 46 520 283 KB

Integrity

Task attestation

Status
complete
Completed
Aug 16, 2026, 12:13 AM UTC
Manifest SHA-256
146f8ed108…6ce005cf
Aggregate SHA-256
58b36cc52e…7d3b2682
Episode index SHA-256
edcca19dba…aa96f0ce
Validation SHA-256
c2fa5442df…bec99e4d
Artifact inventory SHA-256
835061a62d…c50ec809