FP4 pretraining needs more exposures to learn rare facts
NVFP4 needed 1.32× and 1.52× as many exposures as BF16 to match recall while loss differed by 0.5% and 2%.
Stories per day · 30 days
2 RSS
Relatednvidia3nemotron2open-weights2agentic-ai1agents1bf161fp41hbm1inference1kv-cache1looped-llm1mixture-of-experts1multimodal1pretraining1quantization1
14 stories, newest first · one card per story
Stories tagged #nvfp4
NVIDIA claims codesign lifts agentic throughput per dollar up to 30x
From Blackwell Ultra to Vera Rubin, HBM per GPU rose about 3% while claimed agentic throughput per dollar rose up to 30x.
NVIDIA releases Dory1.0-4B, a looped 4B reasoning model
The BF16 checkpoint uses a 256K context, OpenMDW-1.1, and variable loop counts without retraining.
Nvidia's Nemotron 3.5 Super VL is an open model built to run enterprise AI agents nonstop
Nvidia released open Nemotron 3.5 Super VL, a 120B multimodal model for always-on enterprise agents.
autotrust/JEV-35B-NVFP4 — new model trending #21 on Hugging Face
AutoTrust released JEV-35B-NVFP4, a 4-bit quantization of its 35B decision model that cuts GPU memory by 65%.
autotrust/JEV-35B — new model trending #28 on Hugging Face
AutoTrust released JEV-35B, an open-weights 35B Qwen3.5 decision model scoring 59.73 on Decision Index 0.3.
autotrust/GLM5.3-Flash-E224-DGX-Spark — new model trending #29 on Hugging Face
autotrust released an unofficial pruned NVFP4 GLM-5.3-Flash build that fits two NVIDIA DGX Sparks.
miniTE aligns native NVFP4 reinforcement learning across engines
Authors report a Qwen3-30B-A3B learner on one B300 node at 33,540 tokens/s, 1.39× BF16 and 1.47× naive TE NVFP4.
autotrust/GEV-26B-Decide-NVFP4 — new model trending #30 on Hugging Face
autotrust published GEV-26B-Decide-NVFP4, an NVFP4 expert-quantized 26B model that fits in about 17 GB.
SoloQ: Calibration-Free Quantization for Diffusion Language Models
SoloQ quantizes diffusion language models to 4-bit without calibration, cutting memory up to 2.61x with NVFP4.
TRIAGE: Direction-Aware Mismatch Stabilization of Native NVFP4 Reinforcement Learning
TRIAGE stabilizes native NVFP4 RL on Qwen3 models, matching full precision with up to 2.3x rollout throughput.
Disaggregated Quantization: Specializing LLM Prefill and Decode
Disaggregated quantization tailors LLM prefill and decode formats, raising accuracy and speeding time-to-first-token.
NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut
NVIDIA's Vera Rubin NVL72 debuts in MLPerf Inference v6.1 with up to 3.7x higher throughput than GB300 NVL72 and 99% scaling efficiency at 288 GPUs.
8
nvidia/Qwen3.8-Flash-Next-NVFP4 — new model trending #28 on Hugging Face
NVIDIA released an NVFP4 4-bit quantized build of Alibaba's Qwen3.8-Flash-Next, a 125B-parameter MoE vision-language model, via Model Optimizer.
9