OpenAI released GPT-6 Sol and GPT-6 Luna. Both are live now in Devin Desktop and Devin CLI.
On FrontierCode 1.1, our proprietary benchmark that grades models on real-world engineering tasks based on quality and mergeability, GPT-6 Sol matches GPT-5.6 Sol’s score at 61% lower cost per task, and lands within four points of GPT-6 Astra at less than half its cost. GPT-6 Luna scores above GPT-5.6 Luna at about a quarter of the cost. At under $0.10 per task, it is the cheapest model on the leaderboard.
In our evaluations, the gains show up in how the models work, not just what they cost. GPT-6 Sol reads about 17% less context than GPT-5.6 Sol to reach the same score and rarely leaves a repo with failing tests after its last edit. GPT-6 Luna’s biggest gains are at low and medium reasoning effort, where it writes real regression tests instead of mock-based ones.