Cognition's SWE-2
Notable signalCognition's coding model, 64% cheaper than Fable 5.1
Launched 2mo ago
Signal
55
Velocity
3
What this means
45%Mid-tier finish likely.
Projecting attention of 190 by end of day-1.
Rising steadily
Steady upward velocity sustained.
Low discussion activity despite the attention.
Audience is voting but not engaging — could signal weak product-market fit.
Prediction
Top-5 finish probability
45%
today
Projected end-of-day votes
190range 143–257
Trajectory
rising
Steady upward velocity sustained.
About
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.
AI Summary
Cognition's SWE-2 is a coding model that achieves 50.0% on FrontierCode 1.1 Main, closely competing with Fable 5.1 at 64% lower cost. It reduces operational turns by 58% and costs 81% less than its predecessor, SWE-1.7, while providing improved performance.
Attention & discussion velocity
Performance
Velocity3
Vote pace vs average
Momentum3
Sustained over 6h
Virality3
Spread × engagement
Engagement3
Discussion per unit of attention
Signal sources
3 of 6 active- Launch tracking10 readings since we picked it up
- AI analysisCoding Tools
- Comment intelligencethread not yet analysed
- Outside discussion17 threads found
- Revenue signalsno match on this domain
- Editorial scoringbelow the signal threshold for this pass
A source that found nothing is a measurement. A source that has not run is a gap. Neither means the launch lacks the thing.