What are the hardware requirements for Midnight Miqu 103B V1.5 3.5bpw H6 EXL2?

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 requires approximately 45.6 GB of VRAM and supports a context window of 31K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 and how large is it?

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 is developed by LoneStriker, a model with 103b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Midnight Miqu 103B V1.5 3.5bpw H6 EXL2?

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 by LoneStriker — VRAM 45.6GB, 31K context

Name: Midnight Miqu 103B V1.5 3.5bpw H6 EXL2
Author: LoneStriker

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 is an open-source language model by LoneStriker. Features: 103b LLM, VRAM: 45.6GB, Context: 31K, Quantized, Merged, LLM Explorer Score: 0.11.

Merged Model Conversational Endpoints compatible Exl2 Llama Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/LoneStriker/Midnight-Miqu-103B-v1.5-3.5bpw-h6-exl2

Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 Parameters and Internals

Model Type

text generation

Additional Notes

This is a 103B frankenmerge of Midnight-Miqu-70B-v1.5 with itself. Model is capable of 32K context.

LLM Name	Midnight Miqu 103B V1.5 3.5bpw H6 EXL2
Repository 🤗	https://huggingface.co/LoneStriker/Midnight-Miqu-103B-v1.5-3.5bpw-h6-exl2
Merged Model	Yes
Model Size	103b
Required VRAM	45.6 GB
Updated	2026-07-27
Maintainer	LoneStriker
Model Type	llama
Model Files	8.5 GB: 1-of-6 8.6 GB: 2-of-6 8.5 GB: 3-of-6 8.5 GB: 4-of-6 8.5 GB: 5-of-6 3.0 GB: 6-of-6
Quantization Type	exl2
Model Architecture	LlamaForCausalLM
Context Length	32764
Model Max Length	32764
Transformers Version	4.35.2
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
Vocabulary Size	32000
Torch Data Type	float16

Rank the Midnight Miqu 103B V1.5 3.5bpw H6 EXL2 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 Midnight Miqu 103B V1.5 3.5bpw H6 EXL2

Best Alternatives	Context / RAM	Downloads	Likes
... Dark Miqu 103B 5.0bpw H8 EXL2	31K / 65.1 GB	10	0
...qu 103B V1.5 EXL2 5.0bpw Rpcal	31K / 64.9 GB	9	1
...qu 103B V1.5 EXL2 4.0bpw Rpcal	31K / 52.1 GB	3	1
...qu 103B V1.5 EXL2 3.0bpw Rpcal	31K / 39.3 GB	2	1
...ght Miqu 103B V1.0 EXL2 4.5bpw	31K / 58.7 GB	7	1
...ght Miqu 103B V1.0 EXL2 4.0bpw	31K / 52.2 GB	6	1
...ht Miqu 103B V1.0 EXL2 4.25bpw	31K / 55.3 GB	3	1
...ght Miqu 103B V1.0 EXL2 3.5bpw	31K / 45.7 GB	9	1
...ght Miqu 103B V1.0 EXL2 3.0bpw	31K / 39.2 GB	4	2
...ght Miqu 103B V1.0 EXL2 5.0bpw	31K / 64.9 GB	5	1

Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/Midnight-Miqu-103B-v1.5-3.5bpw-h6-exl2.