# Google and Cambridge's VeriHarness uses the generating model to audit its own agent rollouts and adds about 6 points on long-horizon tasks

The method treats agreement between rollouts as suspect and disagreement as a clue, and the authors have released the code with about 26,000 agent traces.

- Published: 2026-10-08T06:25:29.519Z
- Canonical: https://polylog.news/ai/2026-10-08/google-and-cambridge-s-veriharness-uses-the-generating-model
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Polylog editors](https://polylog.news), [arXiv](https://arxiv.org/pdf/2610.00972), [GitHub (google-research/veriharness)](https://github.com/google-research/veriharness)

Researchers at the University of Cambridge and Google Cloud AI Research have open-sourced VeriHarness, a harness for checking the results of long-running agent tasks. It needs no reference answers or grading rubrics at test time. According…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-08/google-and-cambridge-s-veriharness-uses-the-generating-model (subscription information: https://polylog.news/pricing).