Compare AI models

Specifications
Parameters
16B30B
Context window
32k128k
API pricing
Cheapest input
—$0.30DeepInfra
Cheapest output
—$1.20DeepInfra
—$120ModelRun · bf16
These models have no shared benchmark scores.
Overview
CompanyDeepSeekMeta
Release dateMay 16 2024Aug 10 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V2 Lite and Muse Glimmer don't publish scores on any of the same benchmarks, so there's no direct head-to-head comparison. DeepSeek V2 Lite shipped 816 days before Muse Glimmer, so benchmark comparisons should account for the intervening progress.

DeepSeek V2 Lite has 16B parameters, while Muse Glimmer has 30B. Context windows are 32k (DeepSeek V2 Lite) vs 128k (Muse Glimmer).

Direct benchmark comparisons are unavailable — DeepSeek V2 Lite and Muse Glimmer don't publish scores on any of the same benchmarks.