Compare AI models

Specifications
Parameters
16B340B
Context window
32k4k
These models have no shared benchmark scores.
Overview
CompanyDeepSeekNVIDIA
Release dateMay 16 2024Jun 14 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2 Lite and Nemotron-4 340B don't publish scores on any of the same benchmarks, so there's no direct head-to-head comparison. DeepSeek V2 Lite shipped 29 days before Nemotron-4 340B, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while Nemotron-4 340B has 340B. Context windows are 32k (DeepSeek V2 Lite) vs 4k (Nemotron-4 340B).

Direct benchmark comparisons are unavailable — DeepSeek V2 Lite and Nemotron-4 340B don't publish scores on any of the same benchmarks.