Nova Micro 1.0

Nova Micro 1.0 - Text-only low-latency model

Nova Micro 1.0 is Amazon’s smallest Nova model, positioned for ultra-low-latency text inference at very low cost.

Defining Characteristics

• Text-only model: The model is documented as text in, text out, rather than multimodal.
• 128K-token context window: AWS and provider listings specify a 128k-token context length.
• Lowest latency in the Nova family: Amazon positions Nova Micro as the lowest-latency member of the Nova family.
• Lightweight task fit: Official and provider descriptions call out summarization, translation, content classification, interactive chat, and brainstorming as suitable uses.

Core Design

• The model is tuned for cost-efficient text inference and is not intended for reasoning-heavy or multimodal tasks.
• Some documentation mentions basic mathematical reasoning and coding abilities, but not as the model’s main focus.