Gemini 4 Hydrogen is now available in Devin

Devin added Gemini 4 Hydrogen to Devin Desktop and Devin CLI. On Devin's FrontierCode 2.0 benchmark, which scores models on real-world engineering tasks for quality and mergeability, it gets 38.2%. That's within a point of GPT-6.1 Sol's 39.0% and close to Gemini 4 Carbon's 40.0%.

The bigger change is cost. Devin says Hydrogen reaches that score at $2.40 per task, about 40% less than Carbon's $4.04. At medium reasoning effort, it already beats Gemini 4 Argon's best result at less than a seventh of the cost per task.

These are Devin's own benchmark numbers. Cost is the mean spend per rollout at list API pricing, including public discounts, and solutions that fail blocking criteria score 0.