Calvin
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Calvin ABCD→D
99.4Task 1 Completion Rate
140
CALVIN ABCD->D
0.4Avg Length
139
Calvin ABC->D
98.1Success Rate (1)
72
Calvin ABC-D
100Task-1 Score
71
CALVIN
99.8Success Rate (1 task)
63
CALVIN (ABC->D)
4.75Average Successful Length
62
CALVIN ABC-D
0.27Average Trajectory Length
48
CALVIN D→D
4.52Average Length
40
CALVIN ABC→D (unseen environment D)
98.5Success Rate (Length 1)
29
CALVIN ABC→D
99.6Success Rate (1 Task)
22
Calvin ABC→D
4.43Average Sequence Length
18
CALVIN ABC->D 1.0
96.8Success Rate (1 Inst)
18
CALVIN long-horizon
96.9Success Rate 1
17
CALVIN
98.4Success Rate (Step 5/5)
16
CALVIN ABC→D Zero-shot
98.8Task 1 Success Rate
16
CALVIN
96.3Task Completion Rate (1)
15
CALVIN
93.8Success Rate (1 Task)
15
CALVIN
2.55Average Length
13
CALVIN ABC→D (test)
96Seq Success (1)
13
Calvin Unconstrained
100SR (3/5)
12
Calvin Causal
100Success Rate (3/5)
12
CALVIN MTLC
95Success Rate
12
CALVIN D→D long-horizon
99.5Success Rate (LH-1)
11
CALVIN 10% ABCD → D
84.1Success Rate (L=1)
11
CALVIN
100Success Rate
10