AIWHYS

Models

SWE-2

2 synchronized stories carry this tag.

Agents & toolsShipped

Devin Fusion brings paired coding agents to Desktop and CLI

Cognition made Fusion available in Devin Desktop and CLI. A frontier lead model plans and reviews while a lower-cost sidekick explores, implements and tests in a separate persistent context; users select both models, with Cognition recommending Fable 5.1 plus SWE-2.

Why it matters: The release moves a two-agent coding harness from Devin Cloud onto local developer surfaces. Delegation can reduce frontier-model spend, but it creates two contexts, two tool users and a review handoff that must be evaluated as one system.

Read storyOriginal source · cognition.com
Frontier modelsShipped

Cognition releases SWE-2 across Devin's local coding tools

Cognition released SWE-2, a coding model post-trained from Kimi K3 with several reasoning-effort levels. It is available in Devin Desktop and CLI, with rollout to Devin web and Fusion; Cognition prices low, medium and high effort at $0.30, $0.75 and $1.50 per million tokens respectively.

Why it matters: A lower-cost coding model now sits inside local and paired-agent workflows where completed-task cost depends on turns, tool calls and corrections rather than list price alone.

Read storyOriginal source · cognition.com