# Qwen3.8-27B-NVFP4 by Inferact Inc.: Open-Weight Model Source: https://savrn.com/models/qwen3-8-27b-nvfp4-3 Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Qwen3.8-27B-NVFP4 (17.6B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 35.3 GB | 42.3 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 17.6 GB | 21.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 8.8 GB | 10.6 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 23, 2026. [Qwen3.8-27B-NVFP4 on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/qwen3-8-27b-nvfp4-3/gpus) ## SAVRN's Notes on Qwen3.8-27B-NVFP4 A second third-party NVFP4 quantization of Qwen3.8-27B: 26.4 GB published, 17.6B parameters by the hub record, Apache 2.0, 262,144-token context. RadixArk's NVFP4 build of the same base is 21.9 GB — the packages differ by 4.5 GB, so these are not the same recipe republished. The architecture table prices the underlying model at 42.3 GB at 16-bit and 10.6 GB at 4-bit; one MI300X at $1.85 an hour covers serving either way. Neither quant ships evaluations on our record, so the choice is empirical: benchmark both against the FP16 original on your workload and keep the one inside your accuracy budget. The surrounding rail is crowded — uncensored derivatives and EXL2/EXL3 repacks of Qwen3.8-27B are listed nearby — useful if you want a specific serving runtime, but provenance vetting falls to you. For most teams the rule is simple: smallest file that passes your eval. SAVRN Research ## Model Card By Inferact Inc., published under apache-2.0, revision 6128240ebaf4. Quantized version of https://huggingface.co/Qwen/Qwen3.8-27B Read Inferact Inc.'s full model card Quantized version of https://huggingface.co/Qwen/Qwen3.8-27B ## Configuration Architecture Qwen3_5ForConditionalGeneration Context length (tokens) 262,144 Layers 64 Hidden size 5,120 Feed-forward size 17,408 Attention heads 24 Key/value heads 4 Head dimension 256 Vocabulary size 248,320 Model type qwen3_5 Quantization modelopt ## Identity and Version Repository Inferact/Qwen3.8-27B-NVFP4 Publisher Inferact Inc. Task Image and text to text Modality Image and text Library transformers Parameters 17.6B parameters Languages Not stated by the source Revision 6128240ebaf4eaa7bad2b3d1c72c37d677c5f462 First published 2026-08-13 Last updated 2026-08-14 ## Files and Weights 23 files, 26.4 GB in total. The weights are 7 files totalling 26.4 GB in safetensors. Weights7 files · 26.4 GB Configuration6 files · 241.0 KB Tokenizer4 files · 22.9 MB Documentation2 files · 11.5 KB Other3 files · 11.3 KB Repository1 file · 1.6 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-00006.safetensors | Weights | 4.0 GB | 4b71c5d9d502 | | model-00002-of-00006.safetensors | Weights | 4.1 GB | d0ceb042ef00 | | model-00003-of-00006.safetensors | Weights | 4.0 GB | b633a26c1d3b | | model-00004-of-00006.safetensors | Weights | 4.0 GB | 485ea312626e | | model-00005-of-00006.safetensors | Weights | 4.0 GB | 2aa61bab1ed2 | | model-00006-of-00006.safetensors | Weights | 5.4 GB | 8214cd21a468 | | nvfp4_experts_mtp.safetensors | Weights | 849.4 MB | c2b7cb6987e4 | | config.json | Configuration | 19.3 KB | — | | generation_config.json | Configuration | 214 B | — | | hf_quant_config.json | Configuration | 15.1 KB | — | | model.safetensors.index.json | Configuration | 205.6 KB | — | | preprocessor_config.json | Configuration | 390 B | — | | video_preprocessor_config.json | Configuration | 385 B | — | | LICENSE | Documentation | 11.3 KB | — | | README.md | Documentation | 182 B | — | | chat_template.jinja | Other | 9.0 KB | — | | crc32.txt | Other | 2.4 KB | — | | safetensors-md5sum.txt | Other | | — | | .gitattributes | Repository | 1.6 KB | — | | merges.txt | Tokenizer | 3.4 MB | — | | tokenizer.json | Tokenizer | 12.8 MB | 0997f410c57a | | tokenizer_config.json | Tokenizer | 17.9 KB | — | | vocab.json | Tokenizer | 6.7 MB | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 26.4 GB [Download from Inferact Inc.](https://huggingface.co/Inferact/Qwen3.8-27B-NVFP4) Released by Inferact Inc. through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Derived from [Qwen/Qwen3.8-27B](https://savrn.com/models/qwen3-8-27b) - Quantized from [Qwen/Qwen3.8-27B](https://savrn.com/models/qwen3-8-27b) ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 26.4 GB | | 16-bit | 35.3 GB | | 8-bit | 17.6 GB | | 4-bit | 8.8 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Questions About Qwen3.8-27B-NVFP4 ### How much GPU memory does Qwen3.8-27B-NVFP4 need? About 42.3 GB at 16-bit and 10.6 GB at 4-bit: the weights (17.6B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Qwen3.8-27B-NVFP4 on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use Qwen3.8-27B-NVFP4 commercially? Yes. Qwen3.8-27B-NVFP4 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors. ### What is Qwen3.8-27B-NVFP4's context length? 262,144 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Image and text to text ### [Qwen3.8-27B-NVFP4](https://savrn.com/models/qwen3-8-27b-nvfp4-2) [RadixArk](https://savrn.com/model-publishers/radixark) The RadixArk Qwen3.8-27B-NVFP4 model is the quantized version of Qwen/Qwen3.8-27B. The quantization was produced at RadixArk using NVIDIA Model Optimizer, following a mixed NVFP4 W4A4 recipe. Run on SGLang: launch command and per-platform recipes in the Qwen3.8-27B cookbook. This model is not owned or developed by RadixArk. It is a quantized derivative of Qwen's model; see the upstream Qwen3.8-27B model card for the source model's capabilities, training information, limitations, and license. Global Developers looking to deploy an off-the-shelf, pre-quantized model in AI agent systems, chatbots, RAG systems, and other AI-powered applications. Hugging Face 08/14/2026 via… Open weights apache-2.0 18.2B parameters 262,144 tokens Model Optimizer [View model](https://savrn.com/models/qwen3-8-27b-nvfp4-2) Model · Image and text to text ### [Swift-Qwen3.8-27B-Uncensored-NVFP4](https://savrn.com/models/swift-qwen3-8-27b-uncensored-nvfp4) [AJ Gazin](https://savrn.com/model-publishers/ajgazin) NVFP4 checkpoint of an abliterated Swift-Qwen3.8-27B (UkisAI's reasoning-efficient fine-tune of Qwen3.8-27B). For vLLM and SGLang. GGUFs for llama.cpp: The Swift 1.5 version is source). - Swift's own NVFP4 recipe, unmodified, from ukisai/Swift-Qwen3.8-27B-NVFP4, calibrated with NVIDIA ModelOpt. - MTP head and vision tower in BF16, bit-identical to the source. 21.9 GB, NVIDIA ModelOpt mixed-precision format. Needs a vLLM with ModelOpt mixed-precision support (tested on 0.29.0). No --quantization flag. Sampling, as for Swift and Qwen: temperature 1.0, topp 0.95, topk 20, minp 0. The model thinks before answering by default. Tested on an RTX 5090 (32 GB) with vLLM 0.29.0: NVFP4 layers on… Open weights other 18.2B parameters 262,144 tokens vllm [View model](https://savrn.com/models/swift-qwen3-8-27b-uncensored-nvfp4) Model · Image and text to text ### [Swift-1.5-Qwen3.8-27B-Uncensored-NVFP4](https://savrn.com/models/swift-1-5-qwen3-8-27b-uncensored-nvfp4) [AJ Gazin](https://savrn.com/model-publishers/ajgazin) NVFP4 checkpoint of an abliterated Swift 1.5 Qwen3.8-27B (UkisAI's reasoning-efficient fine-tune of Qwen3.8-27B). For vLLM and SGLang. GGUFs for llama.cpp: (measured on the BF16 source). - Swift's own NVFP4 recipe, unmodified, from ukisai/Swift-Qwen3.8-27B-NVFP4, calibrated with NVIDIA ModelOpt. The module split matches UkisAI's Swift 1.5 NVFP4 exactly. - MTP head and vision tower in BF16, bit-identical to the source. 21.9 GB, NVIDIA ModelOpt mixed-precision format. Needs a vLLM with ModelOpt mixed-precision support. No --quantization flag. Sampling, as for Swift and Qwen: temperature 1.0, topp 0.95, topk 20, minp 0. The model thinks before answering by default. Same format, recipe, module… Open weights other 18.2B parameters 262,144 tokens vllm [View model](https://savrn.com/models/swift-1-5-qwen3-8-27b-uncensored-nvfp4) Model · Image and text to text ### [Qwen3.8-27B-Uncensored-NVFP4-v100-skinny](https://savrn.com/models/qwen3-8-27b-uncensored-nvfp4-v100-skinny) [Tim Eastwood](https://savrn.com/model-publishers/tangles) Mixed-precision quantization of prepared for dnv2003/v100-skinny. - MLP gateproj, upproj, downproj, and lmhead: NVFP4, group size 16 87c9f8cf83021957d1a1a575c90c9a4eaaf7ef0c See quantization-audit.json and hfquantconfig.json for the complete machine-readable layout. This model has had safety alignment substantially removed. Use it only for lawful, controlled research and add appropriate safeguards before deployment. Open weights apache-2.0 18.2B parameters 262,144 tokens Model Optimizer [View model](https://savrn.com/models/qwen3-8-27b-uncensored-nvfp4-v100-skinny) Model · Image and text to text ### [GLM-5.3-Flash-NestQuant-1.5-4bit](https://savrn.com/models/glm-5-3-flash-nestquant-1-5-4bit) [Jarrel Seah](https://savrn.com/model-publishers/jarrelscy) Run one of these commands on a separate scorer workstation with SSH access to an idle native ARM64 DGX Spark. The launcher securely prompts for the serving API key, downloads and prepares the pinned model and ARM64 Docker image on the Spark, launches the exact accepted 2,630-slot configuration, opens an authenticated SSH tunnel, installs the pinned official scorer, and runs all 80 tasks twice. Reasoning on — reproduces the canonical 89.2151 run configuration: Reasoning off — reproduces the 86.6870 run configuration: Both options were live-tested against the published Spark service: SSH and host-key validation, authenticated tunnel, endpoint/model check, pinned scorer checkout, 73/73… Open weights mit 16.9B parameters 1,048,576 tokens [View model](https://savrn.com/models/glm-5-3-flash-nestquant-1-5-4bit) Model · Image and text to text ### [gemma-4-31B-it-NVFP4A16](https://savrn.com/models/gemma-4-31b-it-nvfp4a16) [WINSTECHNET_AX](https://savrn.com/model-publishers/axteam) 4-bit quantization of google/gemma-4-31B-it (commit 842da37), produced for efficient vLLM serving. Quantized by axteam. - Scheme — W4A16 — FP4 (E2M1) weights group 16 (FP8 E4M3 scale + per-tensor FP32 global scale), BF16 activations, stored as compressed-tensors nvfp4-pack-quantized. - Kept in BF16 — vision tower, vision projection, token embeddings (tied with lmhead), MoE router, and all norms. These are left unquantized to preserve quality (a small fraction of total parameters). - Tokenizer / chattemplate / processorconfig / generationconfig are byte-identical to the original, including Google's July-2026 corrected chattemplate. config.json is the original plus a quantizationconfig block.… Open weights apache-2.0 18.5B parameters 262,144 tokens transformers [View model](https://savrn.com/models/gemma-4-31b-it-nvfp4a16) ## Inferact Inc. [All models and datasets](https://savrn.com/model-publishers/inferact) ## Versions - [6128240ebaf4](https://savrn.com/models/qwen3-8-27b-nvfp4-3/versions/6128240ebaf4) · current 2026-10-08 ## Explore More - [All image and text to text models](https://savrn.com/models/tasks/image-and-text-to-text) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-19. - [Hugging Face record](https://huggingface.co/Inferact/Qwen3.8-27B-NVFP4) - [How the hub is built](https://savrn.com/model-hub/methodology)