Amazon's generative rows here are the Nova family. Premier is the expensive reasoning model. Nova 2 Lite is the everyday model and is the one that lists video input. Nova Micro is text-only and the cheapest reviewed Nova. Titan rows are embedding models, not chat.
Flagship · $2.50 in · $12.50 out · 1,000,000 tokens
Guide note · 19 Nov 2026
OpenRouter calls Nova Premier the most capable Amazon multimodal model for complex reasoning, and a teacher for distilling custom models. Text and image in, text out, a million-token window. The feed has no third-party index for it. This guide does not invent one. The row does not list video input.
The hard reasoning job you are already running on Amazon
Stay on Premier when the task is genuinely hard and the rest of the system is on Amazon. There is no index on the feed to compare it with GPT or Claude. Judge it on your prompts.
Nova Premier 1.0 is listed at $2.50 input and $12.50 output per million tokens, with a 1,000,000-token context window.
Teacher for a smaller custom model
The listing names distillation. That is a specific reason to pay Premier's rate: generate the training signal, then serve traffic on a cheaper Nova.
Nova Premier 1.0 is listed at $2.50 input and $12.50 output per million tokens, with a 1,000,000-token context window.
The tasks are editorial illustrations, not measured benchmarks. The prices in them are read from today's feed.
Everyday · $0.30 in · $2.50 out · 1,000,000 tokens
Guide note · 19 Nov 2026
OpenRouter describes Nova 2 Lite as the fast, cost-effective reasoning model for everyday work that can read text, images, and video. It is the reviewed Amazon default, and the Nova model to use when the input is a clip.
Nova Micro is text-only. The listing calls it the lowest-latency Nova, at a very low cost, with a knowledge cutoff of 2024-10-31. Use it for short text. Do not use it for a photo, a video, or a question about this month's news.
For most people, start with Nova Premier 1.0, then drop to the next model on this page when the task is routine. Read the cheaper-model note under each example.
Are the examples benchmarks?
No. They are concrete tasks used to explain fit. The prices next to them are live figures from the feed. The judgment is a guide note.