Dragonfly partner praises Zhipu: Model benchmark performs well, previously misjudged the gap between open source and closed source models BlockBeats News: On June 18th, Haseeb Qureshi, Managing Partner of Dragonfly, stated in a post that he had long believed that the gap between open source and closed source models would widen due to data gaps, hardware gaps, and increased distillation restrictions. However, this judgment was incorrect. Zhipu Z.ai has reached another level, with excellent performance in model benchmark testing and strong feedback from early use. Haseeb also stated that it is running its Hermes Agent on GLM 5.2 through the AskVenice API and will provide feedback on its usage experience in the future. [Original link]