NVIDIA ships FP4 quant of Qwen3.8 Flash Next for local use, claims near-lossless vs BF16
EAccelerate_42 · x · 2026-10-16
NVIDIA released an FP4 quantized version of Qwen3.8 Flash Next on Hugging Face (nvidia/Qwen3.8-Flash-Next-FP4-Local), targeted at local deployment.
- The model card claims it loses almost nothing compared to the BF16 baseline.
- Built for local inference, with support for vision inputs (images/video).
- Ships with reasoning effort configuration: xhigh (default), medium, and low.
The sharer plans to benchmark it independently, so the near-lossless claim remains to be third-party verified.
More from Infra
- Local AI is a routing problem, not a privacy feature: where each task should run — bigdata · 2026-10-16
- Moving agent work from GPUs to CPUs for durable, predictable actions — cramforce · 2026-10-16
- First TPUv7 agentic benchmark on vLLM claims 1.42x perf-per-dollar over GB300 NVL72 — vllm_project · 2026-10-16
- Phison CEO calls CXMT memory scare a 'ghost story', warns of materials shortage — BenBajarin · 2026-10-16
- Ben Bajarin: The Future of Compute Is Fungible as Hyperscalers Design for Workload Flexibility — BenBajarin · 2026-10-16
- AirLLM runs frontier-size models layer-by-layer: 671B on 12GB VRAM — Aiden_Tech_Ai · 2026-10-16