EMLab releases KnowLine open-weight decision models from 0.8B to 31B

EMLab releases seven open-weight KnowLine decision models for RAG, agent harnesses and enterprise workflows, according to a ModelScope post.

The set includes dense models from 0.8B to 31B parameters and two mixture-of-experts options. The weights are licensed under Apache 2.0 and the serving code under MIT.

The post reports that KnowLine-31B-Gen1 scores 71.86 on Jev Decision Index 0.4.1 Full across private tests. It also says KnowLine-35B-A3B-Gen1.5 activates about 3B parameters and has a 47.1 ms median latency on a single RTX PRO 6000. These benchmark figures come from EMLab's reported results as relayed by ModelScope, and no independent evaluation is cited.