# Prime Intellect says web access lets models find benchmark answers online

> Original title: Web access is a gnarly problem for public evaluations, as the answers are accessible to the models.

- Source: Prime Intellect
- Published: 2026-11-03T18:54:48.000Z
- Bad Influence: https://badinfluence.news/items/cyv8k8tbzmt711ee6znl9pa04
- Original: https://x.com/PrimeIntellect/status/2117691984551055414

## Summary

Prime Intellect says web access is a difficult problem for public evaluations because models can find the answers online. It says HLE-Diamond with tool calls uses a bespoke script to detect circumvention, but models still find ways around it, such as by finding traces online.

## Post

Web access is a gnarly problem for public evaluations, as the answers are accessible to the models.

This is also a problem for HLE-Diamond (with tool calls), which has a bespoke script to detect circumvention.

However, models find ways around it, e.g., by finding traces online:

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWOIVx0aIAAstj9.png&mode=thumb&exp=1793923200&sig=a76a4f3242299e36)