What are the hardware requirements for Galactica 6.7B Evol Instruct 70K?

Galactica 6.7B Evol Instruct 70K requires approximately 27.4 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Galactica 6.7B Evol Instruct 70K and how large is it?

Galactica 6.7B Evol Instruct 70K is developed by GeorgiaTechResearchInstitute, a model with 6.7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Galactica 6.7B Evol Instruct 70K perform on standard benchmarks?

Galactica 6.7B Evol Instruct 70K has the following published scores: MMLU 32.96. Compare against reference models on this page or on the LLM Explorer leaderboards.

Galactica 6.7B Evol Instruct 70K by GeorgiaTechResearchInstitute — VRAM 27.4GB, 2K context

Name: Galactica 6.7B Evol Instruct 70K
Author: GeorgiaTechResearchInstitute

Galactica 6.7B Evol Instruct 70K is an open-source language model by GeorgiaTechResearchInstitute. Features: 6.7b LLM, VRAM: 27.4GB, Context: 2K, License: cc-by-nc-4.0, Instruction-Based, LLM Explorer Score: 0.11, Arc: 42.6, HellaSwag: 49.3, MMLU: 33, GSM8K: 0.4.

Arxiv:2304.12244 Alpaca Dataset:victor123/evol instruc... Galactica Instruct Opt Pytorch Region:us Sharded Wizardlm

Model Card on HF 🤗: https://huggingface.co/GeorgiaTechResearchInstitute/galactica-6.7b-evol-instruct-70k

Galactica 6.7B Evol Instruct 70K Parameters and Internals

Model Type

text-generation

Use Cases

Areas:

Research

Applications:

Scientific tasks, Language models applied to scientific domain

Limitations:

Cautions against production use without safeguards

Training Details

Data Sources:

Papers, Textbooks, Scientific websites, Encyclopedias

Data Volume:

106 billion tokens

Methodology:

Fine-tuning on Evol-Instruct-70k dataset

Context Length:

2048

Training Time:

22 hours

Hardware Used:

8 A100 80GB GPUs

Input Output

Input Format:

Instruction-response prompt

Accepted Modalities:

Text

Output Format:

Text-based responses

LLM Name	Galactica 6.7B Evol Instruct 70K
Repository 🤗	https://huggingface.co/GeorgiaTechResearchInstitute/galactica-6.7b-evol-instruct-70k
Model Size	6.7b
Required VRAM	27.4 GB
Updated	2026-07-14
Maintainer	GeorgiaTechResearchInstitute
Model Type	opt
Instruction-Based	Yes
Model Files	10.0 GB: 1-of-3 9.9 GB: 2-of-3 7.5 GB: 3-of-3 0.0 GB
Model Architecture	OPTForCausalLM
License	cc-by-nc-4.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.28.0.dev0
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	50000
Torch Data Type	float32
Activation Function	gelu

Rank the Galactica 6.7B Evol Instruct 70K 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 Galactica 6.7B Evol Instruct 70K

Best Alternatives	Context / RAM	Downloads	Likes
Galactica 6.7B EssayWriter	2K / 13.4 GB	55	5
Opt Flan Iml 6.7B	2K / 26.6 GB	131	1
OPI Galactica 6.7B	2K / 13.4 GB	116	6
Ev Glaive Math	2K / 27.4 GB	6	0
Ev Glaive Math1	2K / 27.4 GB	5	0
Open Instruct Opt 6.7B Tulu	2K / 26.6 GB	43	2
LongForm OPT 6.7B	2K / 13.3 GB	15	4
...PT 6.7B Erebus Instruct Follow	2K / 13.8 GB	11	3

Note: green Score (e.g. "73.2") means that the model is better than GeorgiaTechResearchInstitute/galactica-6.7b-evol-instruct-70k.