Compare AI models

Specifications
Parameters
16B/236B8B
Context window
128k8k
Benchmarks
HumanEval
90.2%36.22%
ARC-Challenge
92.8%64.42%
MMLU
79.2%69.51%
GSM8K
94.9%58.45%
Overview
CompanyDeepSeekNVIDIA
Release dateJun 14 2024Aug 21 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek Coder V2 leads Mistral NeMo Minitron 8B on 4 of the 4 benchmarks they both report (HumanEval, ARC-Challenge, MMLU, GSM8K). DeepSeek Coder V2 shipped 68 days before Mistral NeMo Minitron 8B, so benchmark comparisons should account for the intervening progress.

DeepSeek Coder V2 has 16B/236B parameters, while Mistral NeMo Minitron 8B has 8B. Context windows are 128k (DeepSeek Coder V2) vs 8k (Mistral NeMo Minitron 8B).

On HumanEval, DeepSeek Coder V2 leads at 90.2% vs Mistral NeMo Minitron 8B at 36.22%. On ARC-Challenge, DeepSeek Coder V2 leads at 92.8% vs Mistral NeMo Minitron 8B at 64.42%. On MMLU, DeepSeek Coder V2 leads at 79.2% vs Mistral NeMo Minitron 8B at 69.51%. On GSM8K, DeepSeek Coder V2 leads at 94.9% vs Mistral NeMo Minitron 8B at 58.45%.