Amazon Nova Models by Amazon Web Services (AWS) AI
Last refreshed 2026-06-24. Next refresh: weekly.
Details
Capabilities
About
Amazon Nova is AWS’s portfolio of foundation models for text and multimodal generative AI, available to builders through Amazon Bedrock. Understanding models include Nova Premier, Pro, Lite (text, image, and video in; text out) and text-only Nova Micro; creative models include Nova Canvas (image) and Nova Reel (video); Nova Sonic covers conversational speech. AWS documents the line for chat, RAG, agents, media generation, and voice experiences at enterprise scale.
Decision facts
- Best fit
- audiovision and multimodal workreasoning
- Capability starting point
- Amazon Nova 2 Lite with 1m context and reasoning, JSON / Tool use, structured outputs, and multimodal inputs
- Lowest tracked input
- Amazon Nova Micro · $0.035/1M · Vercel AI Gateway
- Closest related family
- Amazon Titan Text
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 1m context, reasoning, and JSON / Tool use.
Use when the workload needs audio, 1m context, and multimodal inputs.
Use when the workload needs 300k context and multimodal inputs.
Use when the workload needs 500k context and multimodal inputs.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Amazon Nova 2 Lite | Use when the workload needs 1m context, reasoning, and JSON / Tool use. | 2025-12 | 1m contextreasoningJSON / Tool use | Current |
| Amazon Nova 2 Pro | Use when the workload needs 1m context. | 2025-12 | 1m context | Current |
| Amazon Nova 2 Omni | Use when the workload needs audio, 1m context, and multimodal inputs. | 2025-12 | audio1m contextmultimodal inputs | Current |
| Amazon Nova Lite | Use when the workload needs 300k context and multimodal inputs. | 2025-11 | 300k contextmultimodal inputs | Current |
| Amazon Nova Micro | Use when the workload needs 4k context. | 2025-11 | 4k context | Current |
| Amazon Nova Pro | Use when the workload needs 500k context and multimodal inputs. | 2025-11 | 500k contextmultimodal inputs | Current |
Release Timeline
2 release groupsSpecifications(6 models)
| Model | Released | Context | Vision | Multimodal | Reasoning | JSON / Tool use | Structured Outputs |
|---|---|---|---|---|---|---|---|
| Amazon Nova 2 Lite | 2025-12 | 1m | Yes | Yes | Yes | Yes | Yes |
| Amazon Nova 2 Pro | 2025-12 | 1m | No | No | No | No | No |
| Amazon Nova 2 Omni | 2025-12 | 1m | Yes | Yes | No | No | No |
| Amazon Nova Lite | 2025-11 | 300k | Yes | Yes | No | No | No |
| Amazon Nova Micro | 2025-11 | 4k | No | No | No | No | No |
| Amazon Nova Pro | 2025-11 | 500k | Yes | Yes | No | No | No |
Available From(2 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Amazon Nova Micro | Vercel AI Gateway | $0.035 | $0.14 | Serverless |
| Amazon Nova Lite | Vercel AI Gateway | $0.06 | $0.24 | Serverless |
| Amazon Nova 2 Lite | Vercel AI Gateway | $0.3 | $2.5 | Serverless |
| Amazon Nova Pro | Vercel AI Gateway | $0.8 | $3.2 | Serverless |
