aiqbenchInformații despre AI & calcul cuantic

...

Română · beta. Traducerea a fost făcută automat și este încă în curs de revizuire. Citește pagina în engleză

perspectiva aiqbenchProduse AI

Cognition's SWE-2 demonstrates significant efficiency gains in code generation by leveraging reinforcement learning on Moonshot AI's Kimi K3, achieving performance comparable to leading closed models at a fraction of the cost. However, its deployment is strictly tied to the Devin agent ecosystem rather than offering standalone open weights or APIs, which limits immediate independent integration for developers. The model also introduces selectable reasoning-effort levels trained in a single run, optimizing the trade-off between computational expense and task complexity.

Sursă care acoperăMarkTechPost

Ce este aceastaLectura unui model asupra articolului și a acoperirii conexe, scrisă pentru aiqbench. Nu sunt cuvintele publicației.

Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost

MarkTechPost12 septembrie 20264 min de cititArticolul și perspectiva sunt în engleză.

Cognition, the company behind the Devin coding agent, has released SWE-2, its most capable coding model to date. SWE-2 is post-trained with reinforcement learning from Kimi K3, Moonshot AI’s 2.8T-parameter open model. Cognition reports a score of 50.0% on FrontierCode 1.1 Main, within 1 point of Fable 5.1 at 64% lower cost. It is also […] The post Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost appeared first on MarkTechPost.

Cognition, the company behind the Devin coding agent, has released SWE-2, its most capable coding model to date. SWE-2 is post-trained with reinforcement learning from Kimi K3, Moonshot AI’s 2.8T-parameter open model. Cognition reports a score of 50.0% on FrontierCode 1.1 Main, within 1 point of Fable 5.1 at 64% lower cost. It is also Cognition’s first model with selectable reasoning-effort levels, all trained in a single RL run. Is…

Extras afișat conform termenilor RSS ai publicației · 70 cuvinte · textul integral și figurile sunt pe site-ul publicației

Citește pe MarkTechPost ↗Întreabă despre acest articol