ITPEval - miniF2F Statements: leaderboard

Metric: Pass@1 (%; share of the 3,792 directed miniF2F statement pairs whose single translated statement verifies; zero-shot direct translation at temperature 0: given a source file in Lean 4, Isabelle, Rocq or HOL Light and a target prover, emit raw target code, which must verify under the target prover). Source: arxiv.org. Saturation forecast: Around 2029. 5 models tracked.

Top models

#ModelScore
1DeepSeek V4 Pro31.38
2GPT-5.530.67
3Gemini 3.1 Pro (Preview)13.08
4Qwen 3 235B A22B11.18
5Claude Sonnet 4.63.53

Interactive version: theaggregate.ai/benchmark?slug=itpeval-minif2f-statements · How It Works · Data refreshed daily, snapshot 2026-09-29.