# Frontier Labs Harden Models Against Reasoning Extraction and Distillation

Frontier labs will treat their models' hidden reasoning and outputs as assets to defend against rival distillation. Extraction resistance will be trained into each release and reported as a headline metric, account-level enforcement will run at industrial scale, and labs will increasingly share threat intelligence with each other, making anti-distillation a recurring security and competitive category.

- Conviction: 36 / 100 (weakening)
- Horizon: Emerging (watchlist)
- Tracking since: 2026-11-21T00:00:00.000Z
- Last updated: 2026-11-22T06:36:09.478Z
- Canonical: https://polylog.news/ai/trends/frontier-labs-harden-against-reasoning-extraction
- Publisher: Polylog

## Recent score history

- 2026-11-21: 38
- 2026-11-22: 36

## Recent evidence

- [confirms] OpenAI says GPT-6.2 Astra leaks protected reasoning in 0.2% of adaptive extraction attacks, down from 54% on GPT-6 Astra (2026-11-21): OpenAI says GPT-6.2 Astra leaks protected reasoning in 0.2% of adaptive extraction attacks, down from 54% on GPT-6 Astra. The company says it now acts against thousands to tens of thousands of accounts a week and shares indicators of compromise with other model providers.
