Amazon Nova Micro
Text-only model that delivers the lowest latency responses at very low cost.
Amazon Low-latency text generation at minimal cost
Amazon Nova Micro is a text-only foundation model developed by Amazon and made available through Amazon Bedrock. It is part of the Amazon Nova family, which includes understanding models (Nova Pro, Nova Lite, and Nova Micro) as well as creative content generation models. Nova Micro is specifically designed to deliver the lowest latency responses within the Nova lineup at very low cost, making it a practical choice for applications where speed and cost efficiency are priorities.
Because Nova Micro handles text input and output exclusively, it is well suited for tasks such as summarization, classification, question answering, and other text-based workflows where multimodal capabilities are not required. The model supports a 128,000-token context window, allowing it to process long documents or extended conversations in a single request. It can also be fine-tuned on Amazon Bedrock, enabling developers to adapt it to specific domains or use cases.
What Amazon Nova Micro supports
Low Latency Responses
Designed to return text completions faster than other models in the Nova family, making it suitable for real-time or high-throughput applications.
Long Context Window
Supports up to 128,000 tokens per request, enabling processing of long documents, transcripts, or multi-turn conversations in a single call.
Text Generation
Generates coherent text output for tasks such as summarization, classification, question answering, and instruction following using text-only input.
Fine-Tuning Support
Can be fine-tuned on Amazon Bedrock to customize model behavior for specific domains or tasks, adjusting the balance of accuracy and cost.
Agentic Task Support
Supports agentic workflows, allowing the model to be used in multi-step task pipelines and tool-use scenarios within Amazon Bedrock.
Cost-Efficient Inference
Positioned as the lowest-cost option in the Nova understanding model tier, making it practical for large-scale or budget-sensitive deployments.
Ready to build with Amazon Nova Micro?
Get Started FreeBenchmark scores
Scores represent accuracy — the percentage of questions answered correctly on each test.
| Benchmark | What it tests | Score |
|---|---|---|
| MMLU-Pro | Expert knowledge across 14 academic disciplines | 53.1% |
| GPQA Diamond | PhD-level science questions (biology, physics, chemistry) | 35.8% |
| MATH-500 | Undergraduate and competition-level math problems | 70.3% |
| AIME 2024 | American math olympiad problems | 8.0% |
| LiveCodeBench | Real-world coding tasks from recent competitions | 14.0% |
| HLE | Questions that challenge frontier models across many domains | 4.7% |
| SciCode | Scientific research coding and numerical methods | 9.4% |
Common questions about Amazon Nova Micro
What is the context window size for Amazon Nova Micro?
Amazon Nova Micro supports a context window of 128,000 tokens, allowing long documents or extended conversations to be processed in a single request.
Does Amazon Nova Micro support image or video inputs?
No. Amazon Nova Micro is a text-only model and does not accept image, video, or audio inputs. For multimodal input support, Amazon Nova Lite or Nova Pro would be required.
Where is Amazon Nova Micro available?
Amazon Nova Micro is available through Amazon Bedrock, Amazon's managed service for foundation models.
Can Amazon Nova Micro be fine-tuned?
Yes. Amazon Nova Micro supports fine-tuning on Amazon Bedrock, which allows developers to customize the model for specific domains or use cases.
What was the training data cutoff for Amazon Nova Micro?
Based on available metadata, Amazon Nova Micro was released in December 2024. Amazon has not publicly specified an exact training data cutoff date beyond this release timeframe.
Parameters & options
Explore similar models
Amazon Amazon Nova Lite
Very low cost multimodal model that is lightning fast for processing image, video, and text inputs.
Amazon Amazon Nova Pro
Highly capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks.
Amazon Amazon Nova Canvas
Amazon Nova Canvas is a state-of-the-art image generation model that creates professional-grade images from text or images provided in prompts.
OpenAI GPT-4o Mini
Low-cost, low-latency model surpassing GPT-3.5 Turbo in textual intelligence and multimodal reasoning.
Google Gemini 3.1 Pro
Google's frontier reasoning model delivering enhanced software engineering, agentic reliability, and multimodal intelligence across a 1M-token context window.
DeepSeek DeepSeek-V3
General-purpose LLM from DeepSeek.
Start building with Amazon Nova Micro
No API keys required. Create AI-powered workflows with Amazon Nova Micro in minutes — free.