Compare AI models

Specifications
Parameters
7B/13B/70B8B
Context window
4k8k
Benchmarks
HumanEval
29.9%36.22%
ARC-Challenge
57.4%64.42%
HellaSwag
85.3%83.03%
WinoGrande
80.2%80.35%
MMLU
68.9%69.51%
GSM8K
56.8%58.45%
Overview
CompanyMetaNVIDIA
Release dateJul 18 2023Aug 21 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral NeMo Minitron 8B leads LLaMA 2 on 5 of the 6 benchmarks they both report. LLaMA 2 shipped 400 days before Mistral NeMo Minitron 8B, so benchmark comparisons should account for the intervening progress.

LLaMA 2 has 7B/13B/70B parameters, while Mistral NeMo Minitron 8B has 8B. Context windows are 4k (LLaMA 2) vs 8k (Mistral NeMo Minitron 8B).

On HumanEval, Mistral NeMo Minitron 8B leads at 36.22% vs LLaMA 2 at 29.9%. On ARC-Challenge, Mistral NeMo Minitron 8B leads at 64.42% vs LLaMA 2 at 57.4%. On HellaSwag, LLaMA 2 leads at 85.3% vs Mistral NeMo Minitron 8B at 83.03%. On WinoGrande, Mistral NeMo Minitron 8B leads at 80.35% vs LLaMA 2 at 80.2%. On MMLU, Mistral NeMo Minitron 8B leads at 69.51% vs LLaMA 2 at 68.9%. On GSM8K, Mistral NeMo Minitron 8B leads at 58.45% vs LLaMA 2 at 56.8%.