Kimi K3 Debuts at #3 on DeepSWE, Matching Frontier Closed-Source Models
The Kimi K3 model achieved a #3 ranking on the DeepSWE benchmark, demonstrating frontier-level performance comparable to leading closed-source models like Claude Fable and GPT-5.6 Sol. This marks the first time an open-weights model has reached such a high performance tier on serious private evaluations.