Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

UniGeo

Benchmarks

Task NameDataset NameSOTA ResultTrend
Geometry CalculationUniGeo 1.0 (Calculation)
Overall Accuracy67.8
22
Geometry ProvingUniGeo 1.0 (Proving)
Overall Score97.8
15
Automated Theorem ProvingUniGeo Plane Geometry 1.0 (test)
Output Tokens (Thousands)0.12
12
Geometry Problem SolvingUniGeo calculation (test)
Top-1 Accuracy96.9
11
Plane Geometry Problem SolvingUnigeo (test)
Accuracy84.6
9
Theorem ProvingUniGeo
Solved Count10
6
Formal Theorem ProvingUniGeo (10)
Proof Length4.1
6
Geometry problem solvingUniGeo Prv (test)
Accuracy98.5
6
Geometry problem solvingUniGeo CAL (test)
Accuracy75.2
6
Geometric ProvingUniGeo proof part
Accuracy94.4
4
Geometric ProvingUniGeo
Top-1 Acc51.3
4
Showing 11 of 11 rows