Cognition launches SWE-2 Flash: a coding agent model 3x faster at one-fifth the price
Cognition (the company behind Devin) announced SWE-2 Flash on November 4, 2026. It is pitched as a cheaper, faster sibling to SWE-2: about 3x the speed at one-fifth the price,…
By Charlotte Hayes
Cognition (the company behind Devin) announced SWE-2 Flash on November 4, 2026. It is pitched as a cheaper, faster sibling to SWE-2: about 3x the speed at one-fifth the price, aimed at routine agent work like automations, code review, feature development, and documentation updates, rather than uniform benchmark leadership. It is free to use through November 30, 2026, and available in Devin Desktop and CLI. Some details worth noting: – 276B total parameters, 12B active, adapted from Thinking Machines Lab's Inkling-Small, whose full weights were released on July 30, 2026. – Cognition reports big gains over Inkling-Small but mixed results against the larger SWE-2, framing Flash as a cost-and-speed alternative, not a replacement. – The weights are not public yet. Cognition is building a secure sharing program for trusted partners and labs, and says targeted cybersecurity training lifted…
It is part of a fast-moving week in AI, as researchers, builders, and enterprises keep adjusting to what the latest capabilities mean for them.
