What are the hardware requirements for Mpt 30B Instruct GGML?

Mpt 30B Instruct GGML requires approximately 16.9 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Mpt 30B Instruct GGML and how large is it?

Mpt 30B Instruct GGML is developed by TheBloke, a model with 30b 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 Mpt 30B Instruct GGML?

Mpt 30B Instruct GGML is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Mpt 30B Instruct GGML by TheBloke — VRAM 16.9GB

Name: Mpt 30B Instruct GGML
Author: TheBloke

Mpt 30B Instruct GGML is an open-source language model by TheBloke. Features: 30b LLM, VRAM: 16.9GB, License: cc-by-sa-3.0, Quantized, Instruction-Based, LLM Explorer Score: 0.07.

Arxiv:2108.12409 Arxiv:2205.14135 Composer Ggml Instruct Llm-foundry Mosaicml Mpt Quantized Region:us

Model Card on HF 🤗: https://huggingface.co/TheBloke/mpt-30B-instruct-GGML

Mpt 30B Instruct GGML Parameters and Internals

Model Type

instruction-following

Additional Notes

The model uses a modified decoder-only transformer architecture, integrating technologies such as FlashAttention and ALiBi for improved performance.

Training Details

Data Sources:

competition_math, conceptofmind/cot_submix_original/cot_gsm8k, knkarthick/dialogsum, mosaicml/dolly_hhrlhf, duorc, tau/scrolls/qasper, emozilla/quality, scrolls/summ_screen_fd, spider

Data Volume:

2048 tokens during initial training phase, adapted for 8192 tokens

Methodology:

finetuned on instruction following datasets such as Dolly HHRLHF

Context Length:

16384

Training Time:

8 hours

Hardware Used:

72 A100 40GB GPUs

Model Architecture:

Modified decoder-only transformer

Input Output

Input Format:

Prompt template with instruction and response format

Accepted Modalities:

text

Output Format:

Generated text based on given instruction

LLM Name	Mpt 30B Instruct GGML
Repository 🤗	https://huggingface.co/TheBloke/mpt-30B-instruct-GGML
Model Size	30b
Required VRAM	16.9 GB
Updated	2026-08-06
Maintainer	TheBloke
Model Type	mpt
Instruction-Based	Yes
Model Files	16.9 GB 18.7 GB 20.6 GB 22.5 GB 31.8 GB
GGML Quantization	Yes
Quantization Type	ggml
Model Architecture	AutoModel
License	cc-by-sa-3.0

Rank the Mpt 30B Instruct GGML 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 Mpt 30B Instruct GGML

Best Alternatives	Context / RAM	Downloads	Likes
...e Llama 30B Instruct 2048 GGML	0K / 13.7 GB	4	20
Mpt 30B Chat GGML	0K / 16.9 GB	6	73
...e Llama 30B Instruct 2048 GGUF	0K / 13.5 GB	865	4
WizardLM 30B Uncensored GGUF	0K / 13.5 GB	1910	18

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/mpt-30B-instruct-GGML.