What are the hardware requirements for OPT 6.7B Nerybus Mix?

OPT 6.7B Nerybus Mix requires approximately 13.7 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 OPT 6.7B Nerybus Mix and how large is it?

OPT 6.7B Nerybus Mix is developed by KoboldAI, 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 OPT 6.7B Nerybus Mix perform on standard benchmarks?

OPT 6.7B Nerybus Mix has the following published scores: MMLU 24.47. Compare against reference models on this page or on the LLM Explorer leaderboards.

OPT 6.7B Nerybus Mix by KoboldAI — VRAM 13.7GB, 2K context

Name: OPT 6.7B Nerybus Mix
Author: KoboldAI

OPT 6.7B Nerybus Mix is an open-source language model by KoboldAI. Features: 6.7b LLM, VRAM: 13.7GB, Context: 2K, License: other, LLM Explorer Score: 0.09, Arc: 39.2, HellaSwag: 68.6, MMLU: 24.5, GSM8K: 0.8.

Deploy:azure En Opt Pytorch Region:us Sharded

Model Card on HF 🤗: https://huggingface.co/KoboldAI/OPT-6.7B-Nerybus-Mix

OPT 6.7B Nerybus Mix Parameters and Internals

Additional Notes

This experimental model combines the weights of two source models with a parameter-wise 50/50 blend, producing coherent outputs though less impressive than the 2.7B variant.

LLM Name	OPT 6.7B Nerybus Mix
Repository 🤗	https://huggingface.co/KoboldAI/OPT-6.7B-Nerybus-Mix
Model Size	6.7b
Required VRAM	13.7 GB
Updated	2026-08-04
Maintainer	KoboldAI
Model Type	opt
Model Files	3.1 GB: 1-of-5 3.1 GB: 2-of-5 3.1 GB: 3-of-5 3.0 GB: 4-of-5 1.4 GB: 5-of-5
Supported Languages	en
Model Architecture	OPTForCausalLM
License	other
Context Length	2048
Model Max Length	2048
Transformers Version	4.24.0
Tokenizer Class	GPT2Tokenizer
Beginning of Sentence Token	</s>
End of Sentence Token	</s>
Unk Token	</s>
Vocabulary Size	50265
Torch Data Type	float16
Activation Function	relu
Errors	replace

Rank the OPT 6.7B Nerybus Mix 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 OPT 6.7B Nerybus Mix

Best Alternatives	Context / RAM	Downloads	Likes
...7B Gqa Ub 16 Best For KV Cache	2K / 27.4 GB	190	0
...B Qcqa Ub 16 Best For KV Cache	2K / 27.4 GB	94	0
....7B Qcqa Ub 16 Best For Q Loss	2K / 27.4 GB	60	0
Galactica 6.7B ReFT GSM8k	2K / 13.8 GB	58	0
Galactica 6.7B EssayWriter	2K / 13.4 GB	55	5
Opt Flan Iml 6.7B	2K / 26.6 GB	131	1
Opt 6.7B	2K / 13.4 GB	120463	121
...lactica 6.7B Evol Instruct 70K	2K / 27.4 GB	129	19
Galactica 6.7B Finetuned	2K / 27.4 GB	89	34
OPI Galactica 6.7B	2K / 13.4 GB	116	6

Note: green Score (e.g. "73.2") means that the model is better than KoboldAI/OPT-6.7B-Nerybus-Mix.