# Cognition says higher-quality RL tasks drove its latest model gains

> Original title: Our training proceeded with large-scale RL. While we made some tweaks to our RL recipe from SWE-2, most of the gains came from higher-qua...

- Source: Cognition
- Published: 2026-11-04T19:55:25.000Z
- Bad Influence: https://badinfluence.news/items/gu3ni7ffico9c56y99k7g7wr9
- Original: https://x.com/cognition/status/2118069626269831278

## Summary

Cognition says most gains in its latest training run came from higher-quality, more diverse reinforcement learning tasks rather than changes to its RL recipe from SWE-2. The research team audited task instructions, environment setup, and verifiers, and added several new task types.

## Post

Our training proceeded with large-scale RL. While we made some tweaks to our RL recipe from SWE-2, most of the gains came from higher-quality, more diverse RL tasks.

Our research team carefully audited the instructions, environment setup, and verifiers for our tasks, while also curating several new task types.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWTZyIdXAAAIvm_.jpg&mode=thumb&exp=1794009600&sig=7bb895b3724cf0c0)