Qwen3-32B-AWQ

Home/Models/Qwen/Qwen3-32B-AWQ

Qwen/Qwen3-32B-AWQ

kind: awq
quant: awq
family: qwen
modality: text-generation
root model: Qwen/Qwen3-32B
provider: Qwen
license: apache-2.0
language: —
context: 32,768
library: transformers
vRAM (GB): —
size on disk: —
downloads: 2,005,871
likes: 136
created: —
last modified: 2025-05-21T06:17:51+00:00 (2025)
huggingface: https://huggingface.co/Qwen/Qwen3-32B-AWQ
tags: transformerssafetensorsqwen3text-generationconversationalarxiv:2309.00071arxiv:2505.09388base_model:Qwen/Qwen3-32Bbase_model:quantized:Qwen/Qwen3-32Blicense:apache-2.0text-generation-inferenceendpoints_compatible4-bitawqregion:us

benchmark	category	score	conf	evidence
MMLU	knowledge	90.9	low	`\| Thinking \| bf16 \| 74.9 \| 68.4 \| 90.9 \| 81.4 \|`
AIME 2024	math	81.4	low	`\| Thinking \| bf16 \| 74.9 \| 68.4 \| 90.9 \| 81.4 \|`
GPQA	reasoning	68.4	low	`\| Thinking \| bf16 \| 74.9 \| 68.4 \| 90.9 \| 81.4 \|`

title	author	status	comments	type
TemporalMesh Transformer: 29.4 PPL at 48% compute — beats Mamba, new open-source architecture	vigneshwar234	open	0	disc
Quantization config?	kmouratidis	open	0	disc