Cognition's SWE-2
Cognition's coding model, 64% cheaper than Fable 5.1
launched 1d ago
2 strangers, one vote each.
2 strangers swiped on Cognition's SWE-2: 0% toasted it and 100% roasted it. Verdict: burnt to a crisp 💀
Launched on 13 September 2026 from Product Hunt, in the week of 7 September 2026.
Every verdict, in order
2 in total
One vote per person, and no names: the crowd is anonymous on purpose, which is what makes it say what it thinks. A paid seat carries a name because somebody was paid to put one there.
What they say it does
SWE-2 is Cognition's new coding model, post-trained from Kimi K3 with RL that optimizes for cost and capability at the same time. It hits 50.0% on FrontierCode 1.1 Main, within a point of Fable 5.1 at 64% less, and lands within a few points of GPT-6 Astra at a quarter of the cost. Compared to SWE-1.7 it takes 58% fewer turns and costs 81% less while scoring higher. Available now in Devin Desktop and CLI.
Alternatives to Cognition's SWE-2
All Cognition's SWE-2 alternatives →
Founder and want the vote updates? Email hello@roasty.tech from your @producthunt.com address and we'll set it up.