AI Paper Atlas

Research map / Large language models / Agents, code and math

Goedel-Prover-V2: Scaling Formal Theorem Proving with Scaffolded Data Synthesis and Self-Correction

Yong Lin, Shange Tang, Bohan Lyu, Ziran Yang, Jui-Hui Chung, Haoyu Zhao, Lai Jiang, Yihan Geng, Jiawei Ge, Jingruo Sun, Jiayun Wu, Jiri Gesi and 8 more

ICLR 2026 · Large language models › Agents, code and math · approach: Retrieval-augmented formal theorem proving

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add verifier-guided self-correction (2 rounds) (ablation)Goedel-Prover-V2-32B · MiniF2F test pass@32pass@32 (%): 88.1 -> 90.4Improved
Add verifier-guided self-correction (2 rounds) (ablation)Goedel-Prover-V2-8B · MiniF2F test pass@32pass@32 (%): 84.6 -> 86.7Improved
Add verifier-guided self-correction (ablation)Goedel-Prover-V2-32B · PutnamBench pass@32problems solved: 43 -> 57Improved
Remove compiler error messages from the self-correction prompt (ablation)Goedel-Prover-V2 self-correction · MiniF2F pass@32pass@32Hurt
Keep previous chain-of-thought in the correction context (ablation)Goedel-Prover-V2 self-correction · MiniF2F pass@32pass@32Improved
Extend context to 128k and allow 5 revision iterations (ablation)Goedel-Prover-V2-32B self-correction · MiniF2F pass@32pass@32 (%): 90.4 -> 92.7Improved
Train formalizer with reasoning traces and expert iteration (Goedel-Formalizer-V2 vs Kimina-autoformalizer)Formalizer · 300 Omni-math problemsformalization passes: 161 -> 228Improved
More RL steps (ablation)Goedel-Prover-V2 · MiniF2Fpass@1Improved
More RL steps (ablation)Goedel-Prover-V2 vanilla mode · MiniF2Fpass@NNo clear effect
More RL steps (ablation)Goedel-Prover-V2 correction mode · MiniF2Fpass@NImproved
Model averaging with the base model (increasing weight on base model) (ablation)RL checkpoints · MiniF2Fpass@1 and pass@NMixed
Scale from 8B to 32BGoedel-Prover-V2 · MiniF2F pass@32pass@32 (%): 84.6 -> 88.1Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy