Nvidia's Nemotron 3 Ultra becomes the smartest open US model, but China still leads
Nvidia's Nemotron 3 Ultra has claimed the top position among open-source US models according to Artificial Analysis benchmarks, marking a significant milestone in domestic AI capability. The achievement underscores intensifying competition in the open-weights space, where US labs are narrowing the gap with Chinese counterparts. However, the framing that China still leads suggests Chinese models retain overall superiority on key benchmarks, signaling that despite Nvidia's progress, the geopolitical AI race remains competitive and the open-source frontier continues reshaping model distribution and accessibility.
Modelwire context
Analyst takeThe headline frames this as a US-vs-China race, but the more consequential detail is that Nvidia, historically a compute infrastructure company, now holds a top-tier position in the open model rankings itself, competing directly with labs whose hardware it sells to.
This is the third major Nvidia model story in a single day on Modelwire. The RTX Spark coverage from The Decoder on June 1st described Nvidia pushing inference to the edge on consumer Windows hardware, and the Cosmos 3 piece from Hugging Face the same day showed Nvidia open-sourcing a multimodal model for physical AI. Nemotron 3 Ultra adds a third layer: Nvidia is now staking a claim at the frontier of general reasoning benchmarks, not just specialized domains. Taken together, these moves suggest Nvidia is building a vertically integrated position across training hardware, edge inference chips, and the open model stack itself. That is a different competitive threat to frontier labs than anything the chip business alone represented.
Watch whether DeepSeek or Qwen release a direct benchmark response within the next four to six weeks that reclaims the open-weights top spot, which would confirm the lead is narrow and contested rather than structural.
Coverage we drew on
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsNvidia · Nemotron 3 Ultra · Artificial Analysis · The Decoder
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. The full content lives on the-decoder.com. If you’re a publisher and want a different summarization policy for your work, see our takedown page.