# Frontier models struggle to classify genetic variants as harmless or harmful

> Original title: To give you a taste of what it looks like when distillation doesn't work, we asked three frontier models to classify 1,000 genetic varian...

- Source: Lewis Tunstall
- Published: 2026-11-02T16:44:14.000Z
- Bad Influence: https://badinfluence.news/items/xt3uuzpmaorh606ls2wyvdqtr
- Original: https://x.com/_lewtun/status/2117296736976691426

## Summary

Lewis Tunstall says three frontier models were tested on classifying 1,000 genetic variants as harmless or harmful. GPT-6.1 Sol returned bare labels with its reasoning encrypted, Gemini 4 Argon refused more than 50% of prompts, and Opus 5.5 ran out of tokens on about 20%. He adds that open-weight models performed barely better than random.

## Post

To give you a taste of what it looks like when distillation doesn't work, we asked three frontier models to classify 1,000 genetic variants: harmless or harmful?

GPT-6.1 Sol answered every one with a bare label, its reasoning encrypted. Gemini 4 Argon refused >50% of the time, while Opus 5.5 ran out of tokens on ~20% of prompts.

Open weight models were also terrible at these DNA understanding tasks, barely better than random.

tl;dr almost nothing to learn from 🫠

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWIiFmMWcAEkg96.png&mode=thumb&exp=1793836800&sig=816c94c8facebdd4)