NVFP4

NVIDIA reports that combined optimizations in TensorRT LLM VisualGen make FLUX 3 Image Klein Base up to 3.46x faster for text-to-image and 4.45x faster for image-to-image at 2048x2048 on a single B200, versus uncompiled BF16. Eight B200 GPUs cut median Base text-to-image latency from 14.90 to 3.00 seconds.

NVIDIA released an NVFP4 quantized checkpoint of DeepSeek AI's DeepSeek-V4.1-Flash-1017 on Hugging Face on 10/30/2026. The checkpoint is about 400 GiB, targets NVIDIA Blackwell GPUs, and matches the MXFP4 source closely on six benchmarks, according to NVIDIA's evaluation.

KRAFTON AI has released Raon-1 mini preview, a Mixture-of-Experts language model trained from scratch with 28.8B total and 4.3B active parameters. It is Apache 2.0 licensed, supports 262,144 tokens of context, and ships with a speculative decoding draft model that KRAFTON says reaches up to 2.18× higher output throughput on a single NVIDIA B300.

NVIDIA has published Nemotron 3.5 Super VL in NVFP4 quantization on Hugging Face: a 120B-parameter, 12B-active hybrid Mamba-2/Transformer MoE with a 1M-token context window and text, image and video input. NVIDIA reports the quantized checkpoint stays close to BF16 on most benchmarks. It fits on one DGX Spark, one H200, two H100s or one GB200.

NVIDIA has published Qwen3.8-Flash-Next-FP4-Local, a 96 GiB NVFP4 quantization of Alibaba's 125B-parameter MoE model, on Hugging Face. NVIDIA reports 91.5% on GPQA Diamond versus 92.2% for BF16. It requires NVIDIA's custom vLLM containers because upstream support has not landed.

NVIDIA has published NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4, a 30-billion-parameter Mixture-of-Experts model with only 3B active parameters, a hybrid Mamba-2/MoE/Attention architecture, and support for up to 1 million tokens of context. The NVFP4-quantized checkpoint is designed to run on a single DGX Spark (GB10) or H100 GPU.