SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness

Papers

arxiv:2609.20519

Published on Sep 17

· Submitted by

Gao Sensen on Sep 18

#2 Paper of the day

· NVIDIA

Upvote

124

Authors:

,

,

Sensen Gao ,

,

,

,

,

,

,

,

,

,

Enze Xie ,

Song Han

Abstract

As coding agents move from supervised code completion to unattended, around-the-clock exploration, their work expands from isolated predictions into long trajectories of reasoning, tool use, and feedback. Token efficiency therefore becomes important for scaling recursive self-improvement. We take an RSI-inspired approach at the harness layer, scaling auto-research loops across increasingly numerous and diverse environments for harness rollouts. At this scale, the process yields reusable improvements that transfer beyond their development setting, moving automated harness discovery toward production-level outcomes. Four mechanisms survive selection and form SoL-Pi, spanning action execution, context compaction, observation handling, and delegated reading. On the 51-task EdgeBench evaluation, SoL-Pi achieves performance comparable to Pi across GPT-5.6 Sol and Opus 5 while reducing recorded token traffic by 44.7-49.0% and API cost by about one third. In other words, estimated hourly savings are \8.75-13.50 relative to native Codex and Claude Code harnesses, and \4.36-5.71 relative to Pi.

View arXiv page View PDF Project pageGitHub 2.9k Add to collection

Community

Sensen02

Paper author Paper submitter 5 days ago

As coding agents move from supervised code completion to unattended, around-the-clock exploration, their work expands from isolated predictions into long trajectories of reasoning, tool use, and feedback. Token efficiency therefore becomes important for scaling recursive self-improvement. We take an RSI-inspired approach at the harness layer, scaling auto-research loops across increasingly numerous and diverse environments for harness rollouts. At this scale, the process yields reusable improvements that transfer beyond their development setting, moving automated harness discovery toward production-level outcomes. Four mechanisms survive selection and form SoL-Pi, spanning action execution, context compaction, observation handling, and delegated reading. On the 51-task EdgeBench evaluation, SoL-Pi achieves performance comparable to Pi across GPT-5.6 Sol and Opus 5 while reducing recorded token traffic by 44.7-49.0% and API cost by about one third. In other words, estimated hourly savings are $8.75-$13.50 relative to native Codex and Claude Code harnesses, and $4.36-$5.71 relative to Pi.

researchstudio-bot

5 days ago

This is an automated message from the ResearchStudio team.

We created an interactive ResearchStudio Reel for this paper. It includes a visual poster, a video, and a blog, all available for download in editable formats.

!Visual poster for this paper(https://hf-mirror.com/buckets/researchstudio-bot/trending-paper-reels-live/resolve/published/v1/jobs/8f0908cc-d626-4236-8606-d0504efa4f14/2609.20519/poster.png)

Open the ResearchStudio Reel →

Download all files from HF Mirror

Please give this comment a thumbs up if you find the Reel helpful!

Want to explore or create Reels for more papers? Visit the ResearchStudio demo.

librarian-bot

4 days ago

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on HF Mirror checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment

Upvote

124

Get this paper in your agent:

hf papers read 2609.20519

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2609.20519 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2609.20519 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2609.20519 in a Space README.md to link it from this page.

Collections including this paper 9

← 返回资讯列表