NatureBench Leaderboard
NatureBench Leaderboard is NatureBench Leaderboard. It is ranked #1195 on the AI Radar, in AI agents, first seen 11 days ago and shared in 1 post (6.1k views).
NatureBench Leaderboard Scientific ML Agent Benchmark Nature Bench Can coding agents match the published SOTA of Nature-family papers? We evaluate frontier coding-agent configurations on 90 Nature-family scientific ML tasks in isolated containers. arXiv GitHub Hugging Face Leaderboard Metric Definitions g SOTA-normalized relative gap, g = dir * (m - m_sota) / |m_sota| ; dir handles metric direction. Surpass-SOTA Share of judge-valid tasks with g > 0.1 . Match-SOTA Share of judge-valid tasks with g >= 0 ; g = 0 matches paper SOTA. Median g (all) Median g across all tasks, assigning g = -1 to judge-invalid and no-score tasks. CR Completion…
What people said about NatureBench Leaderboard on X
Agent 能复现科学，还不会发明科学。 NatureBench 的结果很说明问题：AI Agent 已经能在隐藏原方法、禁止搜网的条件下，对打 Nature 系论文作者的 SOTA——甚至 25.6% 的任务真正超越了已发表结果。 但作者点破了关键： Agent 赢的时候，大多是把科学问题翻译成有监督的机器学习任务，然后靠算力和执行把它做完。 也就是说，它赢的是"复现"——给定问题、给定范式，比作者跑得更快更好。 它还没赢的是"发明"——在没有问题的地方提出新问题，在没有方法的地方造出新方法。 复现考验的是执行力；发明考验的是提问力。 而提问，恰恰是最难被 harness、被外化、被 scale 的那部分能力。
— @qingke_ai, 11 days ago · 31 likes · see the post
Alternatives to NatureBench Leaderboard
- muse.ai — New connectors are live today. Come build with us.
- academy.dair.ai — Chat with Paper
- alphaXiv — τ-BENCH is introduced as a benchmark evaluating AI systems on their ability to construct deployable, customer-facing…
- typesafe.ai — yes - we're hiring!
- Sakana Fugu — One model to command them all
- context.dev — powers the branding, company context, and search.
NatureBench Leaderboard in numbers
- Rank on the AI Radar: #1195 of 1321
- Shared in 1 post by 1 account: @qingke_ai
- 6.1k views on those posts
- First seen 11 days ago, last shared 11 days ago
- Pricing seen by Jev: free
- Market: AI agents
FAQ
What is NatureBench Leaderboard?
NatureBench Leaderboard It was first shared on X 11 days ago and is ranked #1195 on the AI Radar.
Is NatureBench Leaderboard free?
Yes, it is free to use.
Who shared NatureBench Leaderboard?
1 account on X, including @qingke_ai, in 1 post totalling 6.1k views.
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 31.3k posts from 4.7k X accounts over the last 14 days, 1.3k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-19 21:10 UTC. Full method.