Cognition Ships SWE-2, Its First Model With Effort Levels, Pushing the Pareto Curve in a Single RL Run
andrew_n_carr · x · 2026-09-11
Cognition released SWE-2, its first coding model to support effort levels, enabled by significant improvements to its length penalty recipe.
Key result: a single RL run pushes the Pareto curve while preserving its shape — medium effort becomes both cheaper and smarter, while max effort learns to spend more tokens and turns to hit the highest scores. The accompanying chart drew praise from researchers like Andrew Carr.
More from coding & agent
- Stanford's PaperFormalizer turns research papers into Lean-verified versions with author in the loop — StanfordUniversity · 2026-10-27
- Hugging Face TRL makes score centering + masked importance sampling default in async GRPO — _lewtun · 2026-10-27
- AutoResearch agent's weird training curve gets dug into instead of discarded — ZimingLiu11 · 2026-10-27
- Firecrawl offers 2.5x credits to a16z Speedrun founders, unveils $75M Series B — devdigest · 2026-10-27
- Chums: open-source Apple Watch app puts Claude Code and Codex on your wrist — brandon_galang · 2026-10-27
- Open-source Clairic Studio turns pasted text into typed data and working code in 5 languages — dfinke · 2026-10-27