Cognition Launches SWE-2 Model Matching Frontier Evals at 70% Lower Cost, Plus Devin Voice
marvinvonhagen · x · 2026-09-11
Cognition announced SWE-2, its new model that scores on par with recent frontier models on leading evals at up to 70% lower cost, claiming it scaled RL to multiple trillions of parameters with a refined recipe pushing the Pareto curve on both capability and cost. Alongside it comes Devin Voice: call Devin on a landline and it ships code, powered by GPT-Live and SWE-2 — a notable step of the agent company extending into the model layer.
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
- 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
- Matt Pocock shares expand/contract trick: split one-way-door PRs to de-risk big changes — mattpocockuk · 2026-10-27