0

ProofBench

Frontier

Automated theorem proving benchmark

Publisher
Vals AI
Domain
Mathematics
Published
Jun 2026
Official leaderboard
vals.ai/benchmarks/proof_bench

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
Vals AI
Attribution policy →

Top score 78.0% by Claude Opus 5 - 51 models reporting (26 frontier)

Score history

51
0%25%50%75%100%Aug 25Nov 25Feb 26May 26Aug 26GPT-5Claude Sonnet 4.5Gemini 3 Pro PreviewClaude Opus 4.6Claude Opus 4.8Claude Opus 5

Top models

51
ProofBenchBar chart with 21 bars. Highest value: Claude Opus 5 at 78.
21 models

Where it's ranked

1

FAQ

What is ProofBench?
Automated theorem proving benchmark
What is the current top score on ProofBench?
The top reported score is 78.0% by Claude Opus 5, across 51 models reporting (26 from frontier labs).