NVIDIA says Reflection AI’s first open-weight model, Beam, was trained on its GB300 NVL72 system and delivers more than 70,000 tokens per second per GPU on Blac
NVIDIA says Reflection AI’s first open-weight model, Beam, was trained on its GB300 NVL72 system and delivers more than 70,000 tokens per second per GPU on Blackwell Ultra from launch. Why it matters That gives operators a concrete, eye-cat
Open discussion →