Reported. MarkTechPost says vLLM Semantic Router released d3-ultra (~27B / 26.09B), an Apache-2.0 open-weight multimodal decision model that reads text, images, and video once and returns probabilities for every answer, scoring 87.2 on the team’s SR-Text benchmark. The project has not been cited here with a matching primary release note we can confirm independently.
MarkTechPost reports that vLLM Semantic Router released d3-ultra, a ~27B (26.09B) Apache-2.0 open-weight multimodal decision model that reads text, images, and video once and returns probabilities for every answer, scoring 87.2 on the team’s SR-Text benchmark.
We record what sources show, with dates. We don't predict or score how likely a rumor is.