Introducing SWE-2: Pushing the Pareto Frontier | Cognition
cognition.com · 3,504 words · saved by 1 readers
Today we’re introducing SWE-2, our most advanced coding model yet. SWE-2 delivers highly competitive agentic coding performance across multiple effort…
By The Cognition Team09.10.26 Today we’re introducing SWE-2, our most advanced coding model yet. It pushes the Pareto frontier of capability and cost, achieving 50.0% on FrontierCode 1.1 Main1, within one point of Fable 5.1 while being 64% cheaper. With SWE-2, we scaled RL to the multi-trillion-parameter regime for the first time, building on the SWE-1.72 training infrastructure and recipe. The key addition is an RL algorithm that trains all reasoning-effort levels in a single run, advancing the whole cost–performance frontier. base model end of training The result is our closest model…
saved by
related reading
- SWE-1.7: Frontier Intelligence at a Fraction of the Costcognition.com
- >10x More Efficient Pretraining — Magicmagic.dev
- Composer2.pdfcursor.com
- Introducing SWE-grep and SWE-grep-mini: RL for Multi-Turn, Fast Context Retrieval | Cognitioncognition.ai
- 2502.18449arxiv.org
- FrontierSWEfrontierswe.com
- FrontierSWEfrontierswe.com
- Claude SWE-Bench Performance \ Anthropicanthropic.com
- Training an Agentic Router for Optimal Cost-Performance on SWE Tasks | Applied Computeappliedcompute.com
- DeepSeek-R1arxiv.org
- [2607.00053] SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasksarxiv.org
- Benchmarking Coding Agents on Databricks’ Multi-Million Line Codebasedatabricks.com