PutnamBench - Lean 4, Answer Supplied (pass@1, 10 tool calls): leaderboard
Number of Putnam competition problems reported solved in Lean 4, with numerical answers supplied in the theorem statement where required. The source reports compute budget 'pass@1, 10 tool calls'. Higher is better; counts stay separate from other languages, answer conditions and budgets. Reported corpus revision: f858fc. Reported Lean version(s): 4.22.
Metric: Problems solved; Lean 4, Answer Supplied; reported compute budget: pass@1, 10 tool calls. Source: github.com. 1 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 (ReAct, 10 turns) | 28 |
Interactive version: theaggregate.ai/benchmark?slug=putnambench-lean-4-answer-supplied-pass-1-10-tool-calls · How It Works · Data refreshed daily, snapshot 2026-10-16.