Amazon Nova 2 Lite API Pricing
Amazon Nova 2 Lite is Amazon's price-performance Nova 2 tier for high-volume multimodal apps: text, image and video input with a 1M-token context. On Amazon Bedrock global cross-region inference it lists $0.30/M input, $0.075/M cached input and $2.50/M output. The US Geo profile costs 10% more. Read from the AWS Price List for Bedrock.
Run the numbers.
Live calculator pre-loaded with Nova 2 Lite global cross-region rates. The US Geo profile is 10% higher, and Bedrock caches at most 20K tokens of a Nova prompt, so long prompts will not reach the high cache setting.
Real-world presets.
Codebase-scale migration
Reading 100-page contracts
Ticket triage
Research planning turn
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Amazon Nova 2 Lite Current | $0.30 cache $0.075 | $2.50 | $0.306 agentic 92/8 | 1M | High-volume multimodal on AWS |
| Gemini 3.8 Flash | $0.75 cache $0.075 | $3.75 | $0.481 +57% · promo | 1M | Multimodal with prompt caching |
| GPT-5.4 mini | $0.75 cache $0.075 | $4.50 | $0.541 +77% | 400K | Lower-latency OpenAI subagents |
| Gemini 3.5 Flash-Lite | $0.30 cache $0.03 | $2.50 | $0.272 −11% | 1M | Google's low-cost multimodal tier |
| DeepSeek V4.1 Flash | $0.30 cache $0.006 | $1.20 | $0.15 −2.0× | 1M | Bulk low-cost traffic |
Frequently asked.
Practical Amazon Nova 2 Lite pricing questions, with Bedrock's published rates kept separate from workload assumptions.
Q · 01 What is Amazon Nova 2 Lite priced at? +
$0.30/M input, $0.075/M cached input and $2.50/M output. The US Geo profile is 10% higher on every line. Batch and Flex halve input and output; Priority costs 1.75x. Read from the AWS Price List for Amazon Bedrock (us-east-1) on 2026-10-13.Q · 02 What is the difference between global and Geo pricing? +
$0.30/$2.50). The Geo (US) cross-region profile keeps requests inside one geography for data residency and costs 10% more. The model card lists no single-region inference for Nova 2 Lite in us-east-1. This page uses the global rate as the baseline.Q · 03 Does Nova 2 Lite have a prompt-cache discount? +
$0.075/M, 75% below base input, and AWS lists cache writes at $0.00. Bedrock caches at most 20K tokens of a Nova prompt with a 5-minute TTL, so on long agent prompts most input still bills at the full rate. Caching is not available on the batch API, where input is $0.15/M and output $1.25/M.Q · 04 What context window does it support? +
1M-token context window and up to 64K output tokens, with text, image and video input, per its Bedrock model card.Q · 05 How does it compare with Gemini 2.5 Flash? +
$0.30/$2.50 token price with a deeper cache-read discount ($0.03/M against Nova's $0.075/M), so its blended cost is lower. Nova 2 Lite's edge is native AWS hosting, data residency and Bedrock integration. Since September 18, 2026 Google serves Gemini 2.5 Flash only to accounts that used it before; for new projects it points to Gemini 3.5 Flash-Lite, which lists the same rate card.