Amazon Nova Pro
Highly capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks.
Amazon Multimodal accuracy, speed, and cost balanced
Amazon Nova Pro is a multimodal foundation model developed by Amazon and made available through Amazon Bedrock. It accepts text and vision inputs and is designed to handle a wide range of tasks where accuracy, response speed, and cost-efficiency all need to be balanced together. It is part of the Amazon Nova family, which also includes Nova Lite and Nova Micro, each targeting different points on the capability-cost spectrum. Nova Pro was released in December 2024 and supports a 300,000-token context window.
Nova Pro is particularly suited for agentic workflows and UI actuation, meaning it can be used to build systems that take sequences of actions or interact with interfaces. It supports fine-tuning on Amazon Bedrock, allowing developers to customize the model for specific domains or cost targets. Within the Nova family, Pro occupies the highest capability tier among the understanding models, making it the appropriate choice when tasks require processing both text and images at scale.
What Amazon Nova Pro supports
Long Context Window
Processes up to 300,000 tokens in a single request, enabling analysis of lengthy documents, codebases, or multi-turn conversations without truncation.
Multimodal Input
Accepts both text and image inputs, allowing the model to reason over visual content alongside written instructions or questions.
Agentic Task Execution
Designed to support multi-step agentic workflows and UI actuation, enabling automated sequences of actions within larger systems.
Fine-Tuning Support
Supports text and vision fine-tuning on Amazon Bedrock, allowing developers to adapt the model to specific use cases or optimize for cost and accuracy.
Fast Response Speed
Tagged as FAST in the model catalog, reflecting that Nova Pro is designed to return responses quickly relative to its capability tier.
Bedrock API Access
Available via Amazon Bedrock, providing a managed API endpoint with no need to handle infrastructure or model hosting directly.
Ready to build with Amazon Nova Pro?
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 | 69.1% |
| GPQA Diamond | PhD-level science questions (biology, physics, chemistry) | 49.9% |
| MATH-500 | Undergraduate and competition-level math problems | 78.6% |
| AIME 2024 | American math olympiad problems | 10.7% |
| LiveCodeBench | Real-world coding tasks from recent competitions | 23.3% |
| HLE | Questions that challenge frontier models across many domains | 3.4% |
| SciCode | Scientific research coding and numerical methods | 20.8% |
Common questions about Amazon Nova Pro
What is the context window for Amazon Nova Pro?
Amazon Nova Pro supports a context window of 300,000 tokens, which allows it to process large documents, long conversations, or extensive codebases in a single request.
What input types does Amazon Nova Pro support?
Nova Pro is a multimodal model that accepts both text and image inputs. It is distinct from Nova Micro, which is text-only.
When was Amazon Nova Pro released?
Amazon Nova Pro was released in December 2024, which also corresponds to its training data cutoff date as listed in the model metadata.
Can Amazon Nova Pro be fine-tuned?
Yes. Amazon Nova Pro supports text and vision fine-tuning through Amazon Bedrock, allowing developers to customize the model for specific tasks or cost requirements.
How does Nova Pro relate to the other Amazon Nova models?
Nova Pro is the highest-capability understanding model in the Amazon Nova family. Nova Lite is a lower-cost multimodal option, and Nova Micro is a text-only model optimized for low latency and minimal cost.
What people think about Amazon Nova Pro
The Reddit threads found do not directly discuss Amazon Nova Pro, focusing instead on benchmarks for other models such as GPT-5.2, Qwen 3, and GLM-4.5.
As a result, no community sentiment, praise, or concerns specific to Amazon Nova Pro can be derived from these threads.
GPT-5.2 is the new champion of the Elimination Game benchmark, which tests social reasoning, strategy, and deception in a multi-LLM environment. Claude Opus 4.5 and Gemini 3 Flash Preview also made very strong debuts.
Scores of Qwen 3 235B A22B and Qwen 3 30B A3B on six independent benchmarks
PACT: a new head-to-head negotiation benchmark for LLMs
GLM-4.5 and gpt-oss-120b added to the Elimination Game benchmark
Parameters & options
Explore similar models
Amazon Amazon Nova Micro
Text-only model that delivers the lowest latency responses at very low cost.
Amazon Amazon Nova Lite
Very low cost multimodal model that is lightning fast for processing image, video, and text inputs.
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.
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.
OpenAI GPT‑5.2 Pro
The most capable model series yet for professional knowledge work.
Anthropic Claude 4.6 Sonnet
Anthropic's most capable Sonnet model yet, delivering near-Opus-level intelligence for coding, agents, and computer use with a 1M token context window.
Start building with Amazon Nova Pro
No API keys required. Create AI-powered workflows with Amazon Nova Pro in minutes — free.