What are the hardware requirements for Qwen1.5 MoE A2.7B Chat?

Qwen1.5 MoE A2.7B Chat requires approximately 28.7 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 MoE A2.7B Chat and how large is it?

Qwen1.5 MoE A2.7B Chat is developed by Qwen, a model with 14.3b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Qwen1.5 MoE A2.7B Chat perform on standard benchmarks?

Qwen1.5 MoE A2.7B Chat has the following published scores: MMLU 60.97. Compare against reference models on this page or on the LLM Explorer leaderboards.

Qwen1.5 MoE A2.7B Chat by Qwen — VRAM 28.7GB, 32K context

Name: Qwen1.5 MoE A2.7B Chat
Author: Qwen

Qwen1.5 MoE A2.7B Chat is an open-source language model by Qwen. Features: 14.3b LLM, VRAM: 28.7GB, Context: 32K, License: other, MoE, LLM Explorer Score: 0.24, Arc: 53.7, HellaSwag: 80.5, MMLU: 61, GSM8K: 28.2.

Chat Conversational En Endpoints compatible Moe Qwen2 moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen1.5-MoE-A2.7B-Chat

Qwen1.5 MoE A2.7B Chat Parameters and Internals

Model Type

text generation, chat

Additional Notes

The model has an inference speed 1.74 times that of Qwen1.5-7B.

Supported Languages

en (unknown)

Training Details

Methodology:

Supervised finetuning and direct preference optimization

Model Architecture:

Transformer-based MoE decoder-only

Input Output

Input Format:

Prompt structure with system and user roles for chat

Accepted Modalities:

text

Output Format:

Generated text

Performance Tips:

Use provided hyper-parameters in `generation_config.json` for better performance in certain cases.

LLM Name	Qwen1.5 MoE A2.7B Chat
Repository 🤗	https://huggingface.co/Qwen/Qwen1.5-MoE-A2.7B-Chat
Model Size	14.3b
Required VRAM	28.7 GB
Updated	2026-07-27
Maintainer	Qwen
Model Type	qwen2_moe
Model Files	4.0 GB: 1-of-8 4.0 GB: 2-of-8 4.0 GB: 3-of-8 4.0 GB: 4-of-8 4.0 GB: 5-of-8 4.0 GB: 6-of-8 4.0 GB: 7-of-8 0.7 GB: 8-of-8
Supported Languages	en
Model Architecture	Qwen2MoeForCausalLM
License	other
Context Length	32768
Model Max Length	32768
Transformers Version	4.39.0.dev0
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	151936
Torch Data Type	bfloat16
Errors	replace

Rank the Qwen1.5 MoE A2.7B Chat 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 MoE A2.7B Chat

Best Alternatives	Context / RAM	Downloads	Likes
Qwen1.5 MoE A2.7B	8K / 28.7 GB	223	0
Qwen1.5 MoE A2.7B	8K / 28.7 GB	209172	227
...Aux Free Sft Math7k 1e 3 Gamma	8K / 28.7 GB	6	1
Qwen1.5 MoE Sft Math7k	8K / 57.3 GB	5	1
Qwen1.5 MoE A2.7B Wikihow	8K / 28.7 GB	11	4
...en1.5 MoE A2.7B Chat GPTQ Int4	32K / 8.4 GB	2155	50

Note: green Score (e.g. "73.2") means that the model is better than Qwen/Qwen1.5-MoE-A2.7B-Chat.