Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Tool Calling on WHEN2TOOL Hard
Loading...
-0.8
Delta Accuracy (ΔAcc)
PROBE&PREFILL
-43.024
-32.062
-21.1
-10.138
May 10, 2026
Delta Accuracy (ΔAcc)
Delta TC (ΔTC)
Delta Acc / -Delta TC
Updated 22d ago
Evaluation Results
Method
Method
Links
Delta Accuracy (ΔAcc)
Delta TC (ΔTC)
Delta Acc / -Delta TC
PROBE&PREFILL
model averaging=six mo...
2026.05
-0.8
-0.24
-3.4
Necessary (N)
model averaging=six mo...
2026.05
-1.9
-0.04
-44.4
Sparse (S)
model averaging=six mo...
2026.05
-11.1
-0.35
-31.5
Necessary + Reason-then-Act
model averaging=six mo...
2026.05
-23.2
-0.7
-32.9
Sparse + Reason-then-Act
model averaging=six mo...
2026.05
-29.3
-0.84
-34.7
No Tool + Reason-then-Act
model averaging=six mo...
2026.05
-36.6
-0.94
-39.1
No Tool (X)
model averaging=six mo...
2026.05
-41.4
-0.69
-60.4
Feedback
Search any
task
Search any
task