Exclusive: Paying for frontier AI models buys 4-month head start at 5x the cost
6 min2
Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost
4 min1
Kimi-maker Moonshot AI targets $2 billion in annual revenue
Cognition's SWE-2 achieves 92.8 on Terminal-Bench 2.1
2 min
Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra
15 min3
Kimi K3 (2.8T) at 1 token/s on a MacBook Pro, streamed from four SSDs
The OpenAI Hack Shows the Genie Is Out of the Bottle
in the wild 5 min1