Research reproduction
PaperBench
reproducing the results of an ML research paper end to end
Rankings
Higher is betterPaperBench — frequently asked questions
- What is PaperBench?
- reproducing the results of an ML research paper end to end
- Which AI model scores highest on PaperBench?
- Qwen3.8-Max by Qwen holds the best PaperBench result among tracked models, at 93% (released Aug 3 2026). Higher scores are better on this benchmark.
- What are the top 1 models on PaperBench?
- 1. Qwen3.8-Max (Qwen) — 93%.
- How many models have a published PaperBench score?
- 1 tracked model has a published PaperBench score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.