What are the hardware requirements for Qwen1.5 32B Chat AWQ?

Qwen1.5 32B Chat AWQ requires approximately 21.2 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen1.5 32B Chat AWQ and how large is it?

Qwen1.5 32B Chat AWQ is developed by Qwen, a model with 32b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Qwen1.5 32B Chat AWQ perform on standard benchmarks?

Qwen1.5 32B Chat AWQ has the following published scores: MMLU 74.99, LMSYS ELO 1203. Compare against reference models on this page or on the LLM Explorer leaderboards.

Qwen1.5 32B Chat AWQ by Qwen — VRAM 21.2GB, 32K context

Name: Qwen1.5 32B Chat AWQ
Author: Qwen

Qwen1.5 32B Chat AWQ is an open-source language model by Qwen. Features: 32b LLM, VRAM: 21.2GB, Context: 32K, License: other, Quantized, LLM Explorer Score: 0.3, ELO: 1203, Arc: 66, HellaSwag: 85.5, MMLU: 75, GSM8K: 7.1.

Arxiv:2309.16609 4-bit Awq Chat Conversational En Endpoints compatible Quantized Qwen2 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen1.5-32B-Chat-AWQ

Qwen1.5 32B Chat AWQ Benchmarks

LMSys ELO 1203 vs 1272 (so35) -5.4%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Qwen1.5 32B Chat AWQ Parameters and Internals

Model Type

text-generation, chat

Additional Notes

Stable support of 32K context length for models of all sizes. No need of 'trust_remote_code'.

Supported Languages

multilingual ()

Training Details

Methodology:

supervised finetuning and direct preference optimization

Context Length:

32000

Model Architecture:

Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention.

Input Output

Input Format:

tokenized input using the tokenizer

Accepted Modalities:

text

Output Format:

text response generation

Performance Tips:

If you encounter code switching or other undesirable results, use provided hyper-parameters in 'generation_config.json'.

LLM Name	Qwen1.5 32B Chat AWQ
Repository 🤗	https://huggingface.co/Qwen/Qwen1.5-32B-Chat-AWQ
Base Model(s)	Qwen/Qwen1.5-32B-Chat Qwen/Qwen1.5-32B-Chat
Model Size	32b
Required VRAM	21.2 GB
Updated	2026-07-27
Maintainer	Qwen
Model Type	qwen2
Model Files	5.0 GB: 1-of-5 5.0 GB: 2-of-5 5.0 GB: 3-of-5 4.6 GB: 4-of-5 1.6 GB: 5-of-5
Supported Languages	en
AWQ Quantization	Yes
Quantization Type	awq
Model Architecture	Qwen2ForCausalLM
License	other
Context Length	32768
Model Max Length	32768
Transformers Version	4.38.2
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	152064
Torch Data Type	float16
Errors	replace

Rank the Qwen1.5 32B Chat AWQ capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Qwen1.5 32B Chat AWQ

Best Alternatives	Context / RAM	Downloads	Likes
...epSeek R1 Distill Qwen 32B AWQ	128K / 19.4 GB	13741	38
...ekR1 QwQ SkyT1 32B Preview AWQ	128K / 19.4 GB	5	8
QwQ 32B AWQ	40K / 19.4 GB	31488	135
Qwen2.5 Coder 32B Instruct AWQ	32K / 19.4 GB	1730092	38
Qwen2.5 32B Instruct AWQ	32K / 19.4 GB	737541	102
Openhands Lm 32B V0.1 AWQ	32K / 19.4 GB	25	9
QwQ 32B Preview AWQ	32K / 19.4 GB	39	24
...Logic Mix 3 32B Mlx OptiQ 5bit	128K / 22.3 GB	41	1
...pSeek R1 Distill Qwen 32B 4bit	128K / 18.5 GB	7795	51
Hydraulic Deepseek 16bit	128K / 65.8 GB	52	0