vLLM and AMD release d3-ultra, an open 26.09B multimodal decision model

1. 1:50 AM
   MarkTechPostNews

   vLLM Semantic Router releases d3-ultra, a 27B open multimodal decision model

   AIThe vLLM Semantic Router team releases d3-ultra, a 26.09B-parameter dense multimodal decision model under Apache-2.0 that returns answer probabilities rather than generating text. The team reports 87.2 on its own SR-Text test, ahead of GLiDE V1.2 at 85.1, and 58.2% on SR-Box across 10 domains, behind d3-turbo at 60.0%. The model supports 131,072 tokens of context per question and is available through Transformers, vLLM Semantic Router, a free API, or a Hugging Face Space.

1. 11:23 PM
   vLLM@vllm_projectOfficial

   AMD and vLLM Semantic Router release open multimodal model d3-ultra

   AIAMD and the vLLM Semantic Router team released d3-ultra, an open multimodal decision model trained and served on AMD Instinct GPUs. The model provides per-answer probabilities, visual grounding, and video tracking, with Apache-2.0 weights available on Hugging Face. According to vLLM's background post, d3-ultra is a 27B model that ranks #2 of 230 on the Jev Decision Index text board, 3.03 points above d3.

2. 11:01 PM
   Xunzhuo@XunzhuoLiuX

   d3-ultra, an open 27B multimodal decision model, ranks #2 on Jev index

   AIThe publisher releases d3-ultra, an open 27B multimodal decision model with Apache-2.0 open weights on Hugging Face. It ranks #2 of 230 on the Jev Decision Index text board, 3.03 points above its predecessor d3. The model handles text, images, and video, boxing what it finds and tracking it through time.

   

The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.