Benchmarks / PaperBench

PaperBench

Built by OpenAI · Giulio Starace, Oliver Jaffe, Dane Sherburn, James Aung, et al. · released 2 Apr 2025

An agent gets one of 20 ICML 2024 Spotlight or Oral papers and must replicate it from scratch: understand the contributions, write a codebase, and run the experiments. Rubrics written with the paper authors break each paper into 8,316 gradable items, and an LLM judge scores the result.

Frontier

26%

Average replication score

IterativeAgent o1-high (36h limit) · harness: IterativeAgent

2 Apr 2025 · Source: OpenAI (benchmark maintainers)

26.0 ± 0.3 over 3 runs with a 36-hour limit. Top of the official leaderboard, which has not been updated since launch. OpenAI reporting its own model.

Emerging

The best result is at 20–60% of the ceiling, or at 50–100% of human parity.

See the full leaderboard at OpenAI

PaperBench: Average replication score over time, 3 recorded results. 0%20%40%60%80%100%Mar 2025Mar 2025Mar 2025Apr 2025Apr 2025Apr 2025 BasicAgent claude-3.5-sonnet: 21% (2 Apr 2025) IterativeAgent o1-high (24h limit): 24.4% (2 Apr 2025) IterativeAgent o1-high (36h limit): 26% (2 Apr 2025)
Dots are recorded results; the line is the best result so far; the red dot is the current frontier. Sources for every point are in the table below.
Embed this card

The image names the institutions behind the numbers and links back to this page.

HTML:

<a href="https://canagentswork.com/benchmarks/paperbench/"><img src="https://canagentswork.com/og/benchmarks-paperbench.png" width="600" height="315" alt="PaperBench: the best result is 26% (IterativeAgent o1-high (36h limit), 2 Apr 2025)." loading="lazy"></a>

Markdown:

[![PaperBench: the best result is 26% (IterativeAgent o1-high (36h limit), 2 Apr 2025).](https://canagentswork.com/og/benchmarks-paperbench.png)](https://canagentswork.com/benchmarks/paperbench/)

What it measures

How much of a research paper's empirical results an agent can reproduce with no access to the authors' code. Rubric leaves cover code development, execution, and result match. The headline is the average replication score across the 20 papers. A lighter Code-Dev variant skips execution and grades only code development; its scores are much higher and are not comparable.

Average replication score: Mean rubric score across the 20 papers on the full PaperBench pipeline, judged by SimpleJudge with o3-mini, averaged over 3 runs. Higher is better.

Status compares the frontier with a ceiling of 100%.

Facts

Grain
Project level: whole projects judged by an acceptance standard
Environment
cli, repo
Grading
llm-judge, rubric
Tasks
20
Human reference
ML PhDs (best of 3 attempts) scored 41.4% on a 3-paper subset after 48 hours, against 26.6% for o1 on the same subset.
Contamination
Agents are told not to use the authors' public code, and a monitor checks for blacklisted repositories. The papers and their code are public, so training data may include them.
Reuse
Code and rubrics are in the MIT-licensed openai/frontier-evals repo. Papers belong to their authors. (open-mit)

Limits to keep in mind

  • The official leaderboard has not changed since 2025-04-02. Newer models are reported only in lab posts, usually on the Code-Dev variant with different judges, so they are not comparable. Source
  • Grading uses an LLM judge. The paper's JudgeEval shows the judge agrees with humans imperfectly, and full grading is expensive. Source
  • The human baseline covers only 3 of 20 papers and 8 participants, and models were run under slightly different conditions. Source
  • A Qwen3.8 model card lists PaperBench scores of 64.8 to 93.0 for several models, but those are Code-Dev runs judged by Claude Opus 4.6. Aggregators show them as full PaperBench scores. Source

Recorded results

A selection that shows the frontier over time. The full leaderboard is at the source.

SystemAverage replication scoreDateSource
BasicAgent claude-3.5-sonnet
harness: BasicAgent
21%2 Apr 2025OpenAI · primary
IterativeAgent o1-high (24h limit)
harness: IterativeAgent
24.4%2 Apr 2025OpenAI · primary
IterativeAgent o1-high (36h limit) · frontier
harness: IterativeAgent
26%2 Apr 2025OpenAI · primary

Timeline

  • 2 Apr 2025 — OpenAI releases PaperBench for replicating ICML papers. Source

Where it sits in the atlas

ML and research engineeringScience and research (partial)

Research replication

Last checked 23 Sep 2026 against 5 primary sources. See an error? Tell us.

How to cite

Credit the original work first: PaperBench by OpenAI (https://arxiv.org/abs/2504.01848).

Then, if you used this page:

Can Agents Work. "PaperBench: frontier results and sources." https://canagentswork.com/benchmarks/paperbench/ (accessed 2026-09-24). CC BY 4.0.
@misc{caw-paperbench,
  title        = {{PaperBench: frontier results and sources}},
  author       = {{Can Agents Work}},
  year         = {2026},
  howpublished = {\url{https://canagentswork.com/benchmarks/paperbench/}},
  note         = {Accessed 2026-09-24. CC BY 4.0}
}