// Hacker Noon · 20 January 2026

OPT-175B is Comparable to GPT-3 While Requiring Only 1/7th the Carbon Footprint

OPT introduces a suite of open-access language models from 125M to 175B parameters, replicating GPT-3 performance while using 1/7th the carbon footprint. Researchers gain full transparency via model weights, code, and training logs, enabling reproducible experiments. OPT-175B performs well in NLP an...

Hacker Noon

@hacker-noon · Meta

hackernoon.com

Hacker Noon@hacker-noon