SWE-2 Flash: Cognition’s Faster, Cheaper Coding Agent

Cognition launched SWE-2 Flash on November 4, a 276-billion-parameter mixture-of-experts coding model that runs at 200 tokens per second — three times faster than its SWE-2 flagship — at one-fifth the cost. The model is free for paid Devin users through November 30, with post-trial pricing undisclosed and weights kept private despite an Apache-licensed open base. Cognition reports mixed benchmarks for SWE-2 Flash, leaving developers to weigh the speed and cost gains against unproven coding performance.

Cognition launched SWE-2 Flash on November 4 — a 276-billion-parameter mixture-of-experts coding model that runs at 200 tokens per second, three times faster than its SWE-2 flagship, and costs one-fifth as much. It is available free for paid Devin users through November 30. The benchmarks are mixed, the post-trial pricing is undisclosed, and Cognition is keeping the weights private despite building the model on an Apache-licensed open base. Here is what developers actually need to know. Speed Is Real. Price Is a Ratio. The 3x speed claim holds up architecturally. SWE-2 Flash is a mixture-of-experts model with 276 billion total … The post