Dragonfly Partners Praise SmartSpectra: Model Benchmark Performs Well, Previous Gap Judgment Between Open Source and Closed Source Models Was Incorrect

BlockBeats News, June 18th, Dragonfly General Partner Haseeb Qureshi posted a message stating that he had long believed that the gap between open-source and closed-source models would widen due to data discrepancies, hardware disparities, and increasing distillation constraints. However, this assessment was incorrect. Smart Spectrum Z.ai has reached another level, with its model benchmark testing performing exceptionally well and receiving strong early user feedback.

Haseeb also mentioned that he is currently running his Hermes Agent on GLM 5.2 using the AskVenice API and will provide feedback on the user experience later.