ArtifactquantizationGGUFOpen weightsof Qwen3 Next 80B A3B Thinking

bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF

This is a quantization of Qwen3 Next 80B A3B Thinking, not an independent model. Parameters, benchmarks, prices and lineage are recorded on the canonical model. Open Qwen3 Next 80B A3B Thinking →

data quality50

Updated 3 h ago · first seen 11 Sept 2026

model_01M294XK19W64JKZJBPGJBS2QM

Artifact facts

Quantization format: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Quantization: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Downloads: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
Likes: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 6 h agomedium
Hugging Face repo: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Base model: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Quantized by: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Pipeline tag: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Gated: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
License: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Release date: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Last modified: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Model card: huggingface.co/bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium
Tags: Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

21 quantisation levels in this repository (IQ1_M, IQ1_S, IQ2_M, IQ2_S, IQ2_XS, IQ2_XXS, IQ3_M, IQ3_XS…).

Hardware fit (this packaging)

23 of 37 device × quantization combinations fit.

Estimated hardware fit
Hardware	Quantization	Device memory	Est. memory	Fits
Apple M3 Ultra	4bit	—	46.3 GB est.	Yes
Mac Studio (Apple M5 Ultra)	4bit	—	46.3 GB est.	Yes
Apple M2 Ultra	4bit	—	46.3 GB est.	Yes
Apple M1 Ultra	4bit	—	46.3 GB est.	Yes
Apple M3 Max	4bit	—	46.3 GB est.	Yes
Apple M4 Max	4bit	—	46.3 GB est.	Yes
Mac Studio (Apple M5 Max)	4bit	—	46.3 GB est.	Yes
MacBook Pro (Apple M5 Max)	4bit	—	46.3 GB est.	Yes
Apple M2 Max	4bit	—	46.3 GB est.	Yes
Apple M1 Max	4bit	—	46.3 GB est.	Yes
Apple M4 Pro	4bit	—	46.3 GB est.	Yes
Mac mini (Apple M5 Pro)	4bit	—	46.3 GB est.	Yes
MacBook Pro (Apple M5 Pro)	4bit	—	46.3 GB est.	Yes
NVIDIA A100 80GB	4bit	80 GB	46.3 GB est.	Yes
NVIDIA H100 SXM	4bit	80 GB	46.3 GB est.	Yes
NVIDIA H100 NVL	4bit	94 GB	46.3 GB est.	Yes
NVIDIA DGX Spark	4bit	128 GB	46.3 GB est.	Yes
NVIDIA H200	4bit	141 GB	46.3 GB est.	Yes
NVIDIA H200 NVL	4bit	141 GB	46.3 GB est.	Yes
NVIDIA B200	4bit	180 GB	46.3 GB est.	Yes
AMD Instinct MI300X	4bit	192 GB	46.3 GB est.	Yes
AMD Instinct MI325X	4bit	256 GB	46.3 GB est.	Yes
NVIDIA DGX B200	4bit	1,440 GB	46.3 GB est.	Yes
Apple M3 Pro	4bit	—	46.3 GB est.	No
Apple M1 Pro	4bit	—	46.3 GB est.	No
Apple M2 Pro	4bit	—	46.3 GB est.	No
Apple M4	4bit	—	46.3 GB est.	No
iMac (Apple M4)	4bit	—	46.3 GB est.	No
Mac mini (Apple M6)	4bit	—	46.3 GB est.	No
MacBook Air (Apple M5)	4bit	—	46.3 GB est.	No
MacBook Pro (Apple M5)	4bit	—	46.3 GB est.	No
Apple M2	4bit	—	46.3 GB est.	No
Apple M3	4bit	—	46.3 GB est.	No
Apple M1	4bit	—	46.3 GB est.	No
NVIDIA GeForce RTX 3090	4bit	24 GB	46.3 GB est.	No
NVIDIA GeForce RTX 4090	4bit	24 GB	46.3 GB est.	No
NVIDIA GeForce RTX 5090	4bit	32 GB	46.3 GB est.	No

Assumptions (7)

Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

Timeline

11 Sept 2026
Parameters changedArtifactbartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUFbartowski
bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF: parameter count changed from 80000000000 to 79674391296
Parameters80B→79.7Bhuggingface
huggingface.co
11 Sept 2026
Parameters changedArtifactbartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUFbartowski
bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF: parameter count changed from 79674391296 to 80000000000
Parameters79.7B→80Bhuggingface
huggingface.co

Provenance

Attributed facts

Source tiers

T225

Freshest observation

3 h ago

Conflicts

None

Source documents
Source	Document	Type	Tier	Last observed	Snapshots
Hugging Face Hub (public pages, model cards, papers)	huggingface.co/bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF/raw/main/README.md	model_card	T2· Quality secondary	5 h ago	1
Hugging Face Hub (public pages, model cards, papers)	huggingface.co/bartowski/Qwen_Qwen3-Next-80B-A3B-Thinking-GGUF	model_page	T2· Quality secondary	6 h ago	5
Hugging Face Hub (public pages, model cards, papers)	huggingface.co/models?author=bartowski&p=0&sort=downloads	listing	T2· Quality secondary	6 h ago	6

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.