Kimi K3 对比 Claude Fable 5：成本与编码能力分析
原标题：Kimi K3 vs Claude Fable 5 on DeepSWE: Cost and Coding
AI 摘要
Together AI 发布了对 Kimi K3 与 Claude Fable 5 在 DeepSWE 基准上的对比分析。结果显示，Kimi K3 在 pass@1 上以 68.5% 略低于 Fable 的 69.9%，但在 pass@2 和 pass@4 上反超，且每次运行成本仅为 4.65 美元，远低于 Fable 的 13.41 美元，每美元解决任务数是后者的 2.8 倍。Kimi K3 作为开源模型，提供了更高的性价比和部署灵活性。
正文节选
Kimi K3 matches Claude Fable 5 on quality, costs a third as much per solved task, and as an open model gives teams full control over their deployment. - Kimi K3 vs Claude Fable 5 is close on DeepSWE pass@1: Fable leads 69.9% to 68.5%, a 1.4 point gap. - Give the models more attempts and Kimi K3 pulls ahead. It wins pass@2 (82.0 vs 80.2) and pass@4 (89.4% vs 88.5%). - Kimi K3 is far cheaper: \$4.65 per rollout vs \$13.41, and 2.8x more solved tasks per dollar. - Claude Fable 5 is the more reliabl