Meta/Dense

OPT 66B

Name: OPT 66B
Author: Meta

Largest OPT. Requires multi-GPU but quality doesn't match modern 13B models.

chat

66B

Parameters

Context length

Benchmarks

Quantizations

40K

HF downloads

Architecture

Dense

Released

2022-05-03

Layers

KV Heads

Head Dim

128

Family

opt

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_M	2.93	24.7 GB	low
Q2_K	3.16	26.6 GB	low
IQ3_XXS	3.25	27.3 GB	low
IQ3_XS	3.5	29.4 GB	low
Q3_K_S	3.64	30.5 GB	low
IQ3_M	3.76	31.5 GB	low
Q3_K_M	4	33.5 GB	low
Q3_K_L	4.3	36.0 GB	moderate
IQ4_XS	4.46	37.3 GB	moderate
Q4_K_S	4.67	39.0 GB	moderate
Q4_K_M	4.89	40.8 GB	good
Q5_K_S	5.57	46.4 GB	good
Q5_K_M	5.7	47.5 GB	good
Q6_K	6.56	54.6 GB	excellent
Q8_0	8.5	70.6 GB	lossless
FP16	16	132.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN OPT 66B NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run opt:66b-q4_K_M

Tag may need adjustment — check ollama.com/library/opt for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware