#reasoning LLM

'NVIDIA releases Nemotron-Elastic-12B, a single 12B checkpoint that contains nested 9B and 6B variants with no extra training, cutting token and storage costs while keeping strong reasoning performance.'

'Baidu released ERNIE-4.5-21B-A3B-Thinking, a 21B-parameter MoE model with 3B active parameters and 128K context length designed for efficient multi-step reasoning.'