NVIDIA says its collaboration with the SGLang and Miles communities spans “100+ day-zero model launches” and more than two years of engineering. The company shared the figures on November 13, 2026, while celebrating its presence at the SGLang Summit. Source
An earlier NVIDIA post that day cited “60+ day-zero model launches”; roughly 50 minutes later, a second post put that figure at “100+.” The text of the two posts doesn’t explain the difference or identify the later tally as a correction. That’s a bigger reported count, not evidence of dozens of launches in 50 minutes. SourceSource
NVIDIA also says the collaboration spans five generations of accelerated computing, from NVIDIA Ampere to Vera Rubin. It counts 10-plus developer events and meetups alongside the engineering work. Source
Neither post’s text lists the models, defines “day-zero model launches” or explains how NVIDIA counted them. It doesn’t specify model support by hardware, either. For developers checking compatibility, that leaves a celebratory headline number—not a compatibility checklist. SourceSource
Comments