开放模型新动态：Inkling、Hy3、Kimi K3 等展示帕累托前沿价值
原标题：Latest open artifacts (#23): Laguna S2.1, Inkling, & Kimi K3 show the utility of open models on the Pareto frontier
AI 摘要
Interconnects AI 发布第23期开放模型盘点，指出开放模型生态持续繁荣，而非此前预测的行业整合。文章重点介绍了 Thinking Machines 的 Inkling、腾讯的 Hy3、Poolside 的 Laguna S2.1、DeepSeek-V4-Flash 和 Moonshot 的 Kimi K3 等模型，并讨论了开放模型的商业价值、许可证变化及潜在政策影响。
正文节选
Latest open artifacts (#23): Laguna S2.1, Inkling, & Kimi K3 show the utility of open models on the Pareto frontier Capacity to train strong models is proliferating. Consolidation has been one of the paths that many astute observers predicted for the near-future of labs training models. It was labelled as inevitable, as training costs are increasing by orders of magnitude every year. Yet, as someone who in 2024 would’ve predicted consolidation really picking up come 2026 or 2027, where are we? W