SWE-2 is live! Really proud of the work our research team put into this one. Our first model where we're seeing performance comparable to the frontier.
Making it free for the next month for anyone on a Devin plan - let us know what you think!
顯示更多
Introducing SWE-2, our closest model yet to the frontier.
On leading evals, it scores on par with recent frontier models – at up to 70% lower cost.
We scaled RL to multiple trillions of parameters, with a refined recipe that pushes the Pareto curve on both capabilities & cost.
顯示更多
