Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Long-horizon Manipulation on RLBench (Block Stacking Subset)
Loading...
88
Bridge Between Towers Success Rate
RoboStream-32B
4.8
26.4
48
69.6
Mar 13, 2026
Bridge Between Towers Success Rate
Cover Top Block with Box Success Rate
Cover Bottom Block with Box Success Rate
Place Blocks in Two Containers Success Rate
Place Blocks in Two Containers (Hard) Success Rate
Stack Five Colors Success Rate
Stack Three Colors Success Rate
Unstack then Stack Four Colors Success Rate
Average Success Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Bridge Between Towers Success Rate
Cover Top Block with Box Success Rate
Cover Bottom Block with Box Success Rate
Place Blocks in Two Containers Success Rate
Place Blocks in Two Containers (Hard) Success Rate
Stack Five Colors Success Rate
Stack Three Colors Success Rate
Unstack then Stack Four Colors Success Rate
Average Success Rate
RoboStream-32B
Model Scale=32B, Zero-...
2026.03
88
84
68
100
76
72
96
76
82.5
RoboStream-235B
Model Scale=235B, Zero...
2026.03
88
92
96
100
88
80
96
84
90.5
RoboStream-8B
Model Scale=8B, Zero-s...
2026.03
76
40
16
100
24
60
96
52
58
SoFar
Number of episodes=25
2026.03
52
4
0
100
4
4
60
0
28
VoxPoser
Number of episodes=25
2026.03
8
4
0
96
4
4
96
0
26.5
Feedback
Search any
task
Search any
task