Today for AI
HOT RADAR
opensourceHEAT 14.3°

AI CURATED HOT STORY · 10/27/2026

vLLM Open-Sources d3-ultra: A 27B Probabilistic Decision Model for Multimodal Single-Pass Inference

1 reports archived1 independent sourcesupdated 10/27/2026, 13:52:46
Synthesis & Latest Updates
1 Sources Cross-Validated

The vLLM Semantic Router team has open-sourced d3-ultra, a 26-billion-parameter probabilistic decision model. Utilizing a 'state once' architecture, it processes input in a single pass to simultaneously handle diverse tasks like multiple-choice questions, bounding box detection, and video tracking, achieving high efficiency on AMD MI325X hardware. While it outperforms competitors on text semantic routing benchmarks, it trails its lighter sibling in box grounding accuracy, with some test sets remaining private.

LATEST/The vLLM Semantic Router team has open-sourced d3-ultra, a 26-billion-parameter probabilistic decision model. Utilizing a 'state once' architecture, it processes input in a single pass to simultaneously handle diverse tasks like multiple-choice questions, bounding box detection, and video tracking, achieving high efficiency on AMD MI325X hardware. While it outperforms competitors on text semantic routing benchmarks, it trails its lighter sibling in box grounding accuracy, with some test sets remaining private.

Hourly heat curve

12 fully observed hours
Now
14.3
Peak
18.014:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. Marktechpost (@Marktechpost)T2·78 pts

    vLLM Open-Sources d3-ultra: A 27B Probabilistic Decision Model for Multimodal Single-Pass Inference

    • d3-ultra is a 26B parameter Apache-2.0 open-source model featuring the 'State Once' mechanism, enabling parallel query processing after a single encoding step.
    • The model outputs probability distributions instead of text, making it suitable for structured tasks like classification, regression, object detection, and video tracking.
    • It scores 87.2 on the SR-Text benchmark, surpassing Fastino GLiDE V1.2 (85.1) and Liquid AI d1-ultra (83.4), though it slightly underperforms d3-turbo in Box Grounding.