TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

developer tools

The Harness Changed the Solver

Keeping model weights fixed, context compaction and stall handling changed coding-agent results.

Published Updated Story ID: mp-2026-08-30-017
Read the complete editionStory JSON

Summary

Keeping model weights fixed, context compaction and stall handling changed coding-agent results.

On a 169-task tight-context SWE-bench Verified cohort, the treatment raised complete solutions from 43 to 72 and improved mean fail-to-pass fraction. Wider-context results were closer, reinforcing that evaluations measure a model-harness pair rather than model weights alone.

Why it matters

Keeping model weights fixed, context compaction and stall handling changed coding-agent results.

Limits and context

No additional limitation was separately recorded.

Key claims

  1. Keeping model weights fixed, context compaction and stall handling changed coding-agent results.

    Evidence: source-2026-08-30-019

Sources

  1. arXiv preprint 2608.26218arXiv · primary research

Corrections

No corrections have been recorded for this story.