Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Task Goal Completion on AppWorld
Loading...
18.46
Average Completion Score @4
HINT-SD-Multi
5.4808
8.8504
12.22
15.5896
May 18, 2026
Average Completion Score @4
Best Completion Score @4
Updated 15d ago
Evaluation Results
Method
Method
Links
Average Completion Score @4
Best Completion Score @4
HINT-SD-Multi
backbone=Qwen3-4B-Inst...
2026.05
18.46
31.11
HINT-SD-Single
backbone=Qwen3-4B-Inst...
2026.05
16.54
29.4
SDPO
backbone=Qwen3-4B-Inst...
2026.05
9.74
19.32
OpenClaw-RL
backbone=Qwen3-4B-Inst...
2026.05
7.65
12.31
GRPO
backbone=Qwen3-4B-Inst...
2026.05
7.49
15.21
SFT
mode=supervised fine-t...
2026.05
6.82
13.16
Initial
mode=zero-shot, backbo...
2026.05
5.98
13.85
Feedback
Search any
task
Search any
task