ZMIME

Aurora2-400M-A20M vs Aion 1.0 Instruct 2.3B

Compare documented capabilities and access. A dash means a comparable figure is unavailable.

Model specifications
SpecificationAurora2-400M-A20MAuroraAI ResearchAion 1.0 Instruct 2.3BMicrosoft
TaskBase language modelingOn-device language generation
Published capabilitiesEnglish base model pretrained on 100 billion tokens with one shared expert and 32 routed experts per layer.Windows-oriented instruction model for local text generation, including document and email tasks.
Model size400M total / approximately 19M active parameters2.3B parameters
AccessPublic Transformers checkpoint for completion and adaptationDownloadable checkpoint
Operating limits—32K-token context window
SourcesModel weightsModel weights
InputsTextText
WeightsOpen weights · Apache-2.0Open weights · MIT
ReleasedNov 18, 2026Nov 18, 2026