ResearchHarness
🚀
1
Lightweight harness for tool-using LLM agents.
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research
Sci-CoE: Co-evolving Scientific Reasoning LLMs via Geometric Consensus with Sparse Supervision
black-yt
updated a
Space 1 day ago
Add task submission Earth_005
1
#11 opened 2 days ago
by 
black-yt
black-yt
updated a
dataset 1 day ago
Improve dataset card: add task category and sample usage
1
#10 opened 4 days ago
by 
nielsr
Leon-AI
updated a
model 6 days ago
Leon-AI
updated a
dataset 6 days ago
Leon-AI
published a
model 6 days ago
black-yt
authored a
paper 9 days ago
Soptq
authored a
paper 9 days ago
Benchmark allow-list request for ResearchClawBench
🔥 1
9
#9 opened 29 days ago
by black-yt
Post
4840
Hey all — our ResearchClawBench leaderboard just updated 🔥
We let AI do real science: 40 tasks across 10 disciplines, compared to human papers. Hard example? 🏔️ Glacier mass change — AI must integrate 233 datasets from 35 teams, 4 methods, reproduce 6542±387 Gt ice loss vs IPCC. No toy problems.
Latest leaderboard (2026-06-09) 📊:
Agents: 🥇 Claude Code 21.5 (50 = match human), $5.3; 🥈 EvoScientist 18.8, $4.1; 🥉 Codex CLI 18.4, just $2.0
LLMs+Harness: 🥇 Claude-Opus-4.8 21.1, $4.0; 🥈 Claude-Opus-4.7 20.7; 🥉 MiniMax-M3 19.8, only $0.45; Qwen3.7-Max 18.7, $0.42, 11min 💥
Claude still king, but MiniMax/Qwen/DeepSeek are crazy cheap and competitive. Expensive isn't always better.
📎 Code & star: https://github.com/InternScience/ResearchClawBench
🏠 Website: https://internscience.github.io/ResearchClawBench-Home/
🤗 Upvote paper: ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research (2606.07591)
We let AI do real science: 40 tasks across 10 disciplines, compared to human papers. Hard example? 🏔️ Glacier mass change — AI must integrate 233 datasets from 35 teams, 4 methods, reproduce 6542±387 Gt ice loss vs IPCC. No toy problems.
Latest leaderboard (2026-06-09) 📊:
Agents: 🥇 Claude Code 21.5 (50 = match human), $5.3; 🥈 EvoScientist 18.8, $4.1; 🥉 Codex CLI 18.4, just $2.0
LLMs+Harness: 🥇 Claude-Opus-4.8 21.1, $4.0; 🥈 Claude-Opus-4.7 20.7; 🥉 MiniMax-M3 19.8, only $0.45; Qwen3.7-Max 18.7, $0.42, 11min 💥
Claude still king, but MiniMax/Qwen/DeepSeek are crazy cheap and competitive. Expensive isn't always better.
📎 Code & star: https://github.com/InternScience/ResearchClawBench
🏠 Website: https://internscience.github.io/ResearchClawBench-Home/
🤗 Upvote paper: ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research (2606.07591)
- 2 replies
black-yt
updated a
Space 9 days ago
black-yt
updated a
collection 9 days ago
black-yt
submitted a
paper to Daily Papers 9 days ago
sY713
authored a
paper 10 days ago
Soptq
authored a
paper 3 months ago
black-yt
authored a
paper 3 months ago
black-yt
authored a
paper 4 months ago
sY713
authored a
paper 4 months ago
sY713
submitted a
paper to Daily Papers 4 months ago