List your company

Autoformalization on miniF2F (test)

96TC@1Oracle*
68.585675.702882.8289.9372Oct 12, 2025

Evaluation results

MethodLinks
Oracle*Formalizer=Goedel-V2-8B2025.1096—————28.8910093.33
RAutoRetrieval Strategy=RAuto, Model=Claude-Opus-42025.1095.98—————30.36——
DPR (RAuto)Formalizer=Goedel-V2-8B2025.1095.55—————9.7710026.67
Zero-shotRetrieval Strategy=Zero-shot, Model=Claude-Opus-42025.1095.09—————31.25——
DRIFTRetrieval Strategy=DRIFT, Model=Claude-Opus-42025.1093.75—————32.59——
Zero-shotFormalizer=Goedel-V2-8B2025.1093.33—————22.2210093.33
DRIFTFormalizer=Goedel-V2-8B2025.1093.33—————26.6710093.33
Oracle*Retrieval Strategy=Oracle*, Model=Claude-Opus-42025.1093.3—————35.27——
Oracle*Formalizer=DeepSeek-V3.12025.1077.68—————23.2187.528.12
DPR (RAuto)Formalizer=GPT-4.12025.1077.23—————24.5592.4132.14
Zero-shotFormalizer=DeepSeek-V3.12025.1076.34—————22.7787.527.23
DPR (RAuto)Formalizer=DeepSeek-V3.12025.1075.89—————22.7787.9527.68
Oracle*Formalizer=GPT-4.12025.1075.45—————23.6689.2930.36
DRIFTFormalizer=GPT-4.12025.1074.55—————24.5592.4129.02
DRIFTFormalizer=DeepSeek-V3.12025.1074.11—————22.7788.8424.55
Zero-shotFormalizer=GPT-4.12025.1069.64—————23.2184.8228.12
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=02026.01—80.3353.5980.1954.9255.69———
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=12026.01—93.4466.686.3470.772.34———
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=22026.01—93.4471.1188.6674.875.98———
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=32026.01—93.4472.3489.4875.8277———
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=42026.01—93.4473.5789.8976.8477.89———
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=52026.01—93.4474.0890.1677.0578.22———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=02026.01—80.3329.9247.2716.827.61———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=12026.01—93.4436.7851.3720.935.03———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=22026.01—93.4439.0452.7322.7536.85———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=32026.01—93.4439.9653.9623.1637.7———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=42026.01—93.4440.0654.5123.7738.13———
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=52026.01—93.4439.8654.6423.9838.17———