Compare AI models

Specifications
Parameters
16B123B
Context window
32k128k
API pricing
Cheapest input
—$2.00Mistral
Cheapest output
—$6.00Mistral
—$700Mistral
Benchmarks
HumanEval
57.3%92.1%
MMLU
55.7%84%
Overview
CompanyDeepSeekMistral
Release dateMay 16 2024Jul 24 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral Large 2 leads DeepSeek V2 Lite on 2 of the 2 benchmarks they both report (HumanEval, MMLU). DeepSeek V2 Lite shipped 69 days before Mistral Large 2, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while Mistral Large 2 has 123B. Context windows are 32k (DeepSeek V2 Lite) vs 128k (Mistral Large 2).

On HumanEval, Mistral Large 2 leads at 92.1% vs DeepSeek V2 Lite at 57.3%. On MMLU, Mistral Large 2 leads at 84% vs DeepSeek V2 Lite at 55.7%.