Skip to content
AI Atlas
ArtifactquantizationBF16/F16/I32/I64Open weightsof Qwen3.8 Flash

Intel/Qwen3.8-Flash-Next-W4A16-AutoRound

published by Intelhuggingface.co/Intel/Qwen3.8-Flash-Next-W4A16-AutoRound

This is a quantization of Qwen3.8 Flash, not an independent model. Parameters, benchmarks, prices and lineage are recorded on the canonical model. Open Qwen3.8 Flash →

data quality42

Updated 3 h ago · first seen 11 Sept 2026

model_01M294YS470NKPRW92DN4R86MW

File size
Format
Downloads
Published

Artifact facts

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Downloads

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 7 h agomedium

Likes

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 7 h agomedium

Hugging Face repo

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 13 h agomedium

Base model

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 13 h agomedium

Library name

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 13 h agomedium

Tags

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Hardware fit (this packaging)

Estimated

23 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
Apple M3 Ultra4bit—43.8 GB est.Yes
Mac Studio (Apple M5 Ultra)4bit—43.8 GB est.Yes
Apple M2 Ultra4bit—43.8 GB est.Yes
Apple M1 Ultra4bit—43.8 GB est.Yes
Apple M3 Max4bit—43.8 GB est.Yes
Apple M4 Max4bit—43.8 GB est.Yes
Mac Studio (Apple M5 Max)4bit—43.8 GB est.Yes
MacBook Pro (Apple M5 Max)4bit—43.8 GB est.Yes
Apple M2 Max4bit—43.8 GB est.Yes
Apple M1 Max4bit—43.8 GB est.Yes
Apple M4 Pro4bit—43.8 GB est.Yes
Mac mini (Apple M5 Pro)4bit—43.8 GB est.Yes
MacBook Pro (Apple M5 Pro)4bit—43.8 GB est.Yes
NVIDIA A100 80GB4bit80 GB43.8 GB est.Yes
NVIDIA H100 SXM4bit80 GB43.8 GB est.Yes
NVIDIA H100 NVL4bit94 GB43.8 GB est.Yes
NVIDIA DGX Spark4bit128 GB43.8 GB est.Yes
NVIDIA H2004bit141 GB43.8 GB est.Yes
NVIDIA H200 NVL4bit141 GB43.8 GB est.Yes
NVIDIA B2004bit180 GB43.8 GB est.Yes
AMD Instinct MI300X4bit192 GB43.8 GB est.Yes
AMD Instinct MI325X4bit256 GB43.8 GB est.Yes
NVIDIA DGX B2004bit1,440 GB43.8 GB est.Yes
Apple M3 Pro4bit—43.8 GB est.No
Apple M1 Pro4bit—43.8 GB est.No
Apple M2 Pro4bit—43.8 GB est.No
Apple M44bit—43.8 GB est.No
iMac (Apple M4)4bit—43.8 GB est.No
Mac mini (Apple M6)4bit—43.8 GB est.No
MacBook Air (Apple M5)4bit—43.8 GB est.No
MacBook Pro (Apple M5)4bit—43.8 GB est.No
Apple M24bit—43.8 GB est.No
Apple M34bit—43.8 GB est.No
Apple M14bit—43.8 GB est.No
NVIDIA GeForce RTX 30904bit24 GB43.8 GB est.No
NVIDIA GeForce RTX 40904bit24 GB43.8 GB est.No
NVIDIA GeForce RTX 50904bit32 GB43.8 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.

Provenance

Attributed facts

23

Source tiers

T223

Freshest observation

3 h ago

Conflicts

None

Source documents
SourceDocumentTypeTierLast observedSnapshots
Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=Intel&p=0&sort=downloads listingT2· Quality secondary3 h ago7
Hugging Face Hub (public pages, model cards, papers)huggingface.co/Intel/Qwen3.8-Flash-Next-W4A16-AutoRound/raw/main/README.md model_cardT2· Quality secondary6 h ago1
Hugging Face Hub (public pages, model cards, papers)huggingface.co/Intel/Qwen3.8-Flash-Next-W4A16-AutoRound model_pageT2· Quality secondary7 h ago4

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