PutnamBench - Isabelle, Answer Supplied (avg $8/problem): leaderboard

Number of Putnam competition problems reported solved in Isabelle, with numerical answers supplied in the theorem statement where required. The source reports compute budget 'avg $8/problem'. Higher is better; counts stay separate from other languages, answer conditions and budgets. A corpus revision is not reported. The source reports that 130 problems require formalization fixes.

Metric: Problems solved; Isabelle, Answer Supplied; reported compute budget: avg $8/problem. Source: github.com. 1 models tracked.

Top models

#ModelScoreOverall rank
1Varto640

Interactive version: theaggregate.ai/benchmark?slug=putnambench-isabelle-answer-supplied-avg-8-problem · How It Works · Data refreshed daily, snapshot 2026-10-16.