Compare AI models

Specifications
Parameters
7B/13B/33B/65B8B
Context window
2k4k
Benchmarks
HellaSwag
84.2%76.4%
WinoGrande
77%70.9%
MMLU
63.4%54.4%
Overview
CompanyMetaNVIDIA
Release dateFeb 24 2023Nov 15 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

LLaMA 1 leads Nemotron-3 8B on 3 of the 3 benchmarks they both report (HellaSwag, WinoGrande, MMLU). LLaMA 1 shipped 264 days before Nemotron-3 8B, so benchmark comparisons should account for the intervening progress.

LLaMA 1 has 7B/13B/33B/65B parameters, while Nemotron-3 8B has 8B. Context windows are 2k (LLaMA 1) vs 4k (Nemotron-3 8B).

On HellaSwag, LLaMA 1 leads at 84.2% vs Nemotron-3 8B at 76.4%. On WinoGrande, LLaMA 1 leads at 77% vs Nemotron-3 8B at 70.9%. On MMLU, LLaMA 1 leads at 63.4% vs Nemotron-3 8B at 54.4%.