AWS Bedrock API Pricing (2026)
Amazon Bedrock is AWS's fully managed service for accessing foundation models from Anthropic, Meta, Mistral, Amazon, and others through a single unified API. It integrates natively with AWS infrastructure, supports VPC endpoints for private connectivity, and offers enterprise features like model customization, guardrails, and AWS IAM-based access control. Founded in 2023 and headquartered in Seattle, WA, AWS Bedrock offers 10 active models through their API. All pricing data below is sourced directly from the AWS Bedrock pricing page and is updated regularly.
Prices verified Mar 11, 2026
AWS Bedrock Model Pricing — All Models
All prices are in USD per 1 million tokens ($/M tokens). Input tokens are the text you send to the model; output tokens are the generated response.
| Model | Tier | Input | Output | Cached Input | Batch Input | Batch Output |
|---|---|---|---|---|---|---|
Claude Sonnet 4.5 (Bedrock) Anthropic's Claude Sonnet 4.5 hosted on AWS Bedrock. Offers strong reasoning and instruction-following with 200K context, delivered through AWS's managed infrastructure with enterprise security and compliance features. | Premium | $3.00/M | $15.00/M | — | — | — |
Claude 3.5 Sonnet v2 (Bedrock) Anthropic's Claude 3.5 Sonnet v2 on AWS Bedrock. Excellent balance of intelligence and speed with 200K context, computer use capability, and strong coding performance, available through AWS's secure managed infrastructure. | Premium | $3.00/M | $15.00/M | — | — | — |
Llama 3.1 70B (Bedrock) Meta's Llama 3.1 70B instruction-tuned model on AWS Bedrock. A capable open-weight model offering strong general performance at mid-range pricing, with 128K context and AWS enterprise integration. | Mid | $1.95/M | $2.56/M | — | — | — |
Llama 3.1 8B (Bedrock) Meta's Llama 3.1 8B instruction-tuned model on AWS Bedrock. A lightweight and cost-efficient model suitable for high-throughput tasks, classification, and simple generation with 128K context. | Budget | $0.22/M | $0.22/M | — | — | — |
Mistral Large (Bedrock) Mistral AI's flagship large model on AWS Bedrock. Offers strong multilingual reasoning, coding, and instruction-following capabilities at premium pricing, accessible through AWS's enterprise infrastructure. | Premium | $4.00/M | $12.00/M | — | — | — |
Amazon Nova Pro Amazon's flagship Nova Pro model, purpose-built for Bedrock. Delivers strong multimodal capabilities including vision, document understanding, and video analysis with a massive 300K context window at competitive mid-range pricing. | Mid | $0.80/M | $3.20/M | — | — | — |
Amazon Nova Lite Amazon's mid-tier Nova Lite model offering multimodal capabilities including text, image, and video understanding at very low cost. Features a 300K context window, ideal for high-volume document and content processing pipelines. | Budget | $0.06/M | $0.24/M | — | — | — |
Amazon Nova Micro Amazon's most cost-efficient Nova model, optimized for text-only tasks at ultra-low latency. Features 128K context and is ideal for high-throughput classification, routing, extraction, and simple generation workloads. | Budget | $0.04/M | $0.14/M | — | — | — |
DeepSeek V3.2 (Bedrock) DeepSeek V3.2 hosted on AWS Bedrock, providing access to DeepSeek's high-performance mixture-of-experts model through AWS's secure infrastructure. Strong coding, math, and reasoning at sub-dollar pricing. | Mid | $0.62/M | $1.85/M | — | — | — |
Gemma 3 27B (Bedrock) Google's Gemma 3 27B instruction-tuned model on AWS Bedrock. A capable open-weight model offering strong multilingual and reasoning performance at budget pricing, accessible through AWS's managed infrastructure. | Budget | $0.23/M | $0.38/M | — | — | — |
Prices last verified: 2026-03-11. Always confirm on the official pricing page before production use.
AWS Bedrock Model Capabilities
Key technical capabilities across all AWS Bedrock models. Use this table to identify which models support the features your application requires.
| Model | Context | Max Output | Vision | Functions | JSON Mode | Caching | Batch API | Reasoning |
|---|---|---|---|---|---|---|---|---|
| Claude Sonnet 4.5 (Bedrock) | 200,000 | 8,192 | ✓ | ✓ | ✓ | — | ✓ | — |
| Claude 3.5 Sonnet v2 (Bedrock) | 200,000 | 8,192 | ✓ | ✓ | ✓ | — | ✓ | — |
| Llama 3.1 70B (Bedrock) | 128,000 | 4,096 | — | ✓ | ✓ | — | ✓ | — |
| Llama 3.1 8B (Bedrock) | 128,000 | 4,096 | — | ✓ | ✓ | — | ✓ | — |
| Mistral Large (Bedrock) | 128,000 | 8,192 | — | ✓ | ✓ | — | ✓ | — |
| Amazon Nova Pro | 300,000 | 5,120 | ✓ | ✓ | ✓ | — | ✓ | — |
| Amazon Nova Lite | 300,000 | 5,120 | ✓ | ✓ | ✓ | — | ✓ | — |
| Amazon Nova Micro | 128,000 | 5,120 | — | ✓ | ✓ | — | ✓ | — |
| DeepSeek V3.2 (Bedrock) | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Gemma 3 27B (Bedrock) | 128,000 | 8,192 | ✓ | ✓ | ✓ | — | — | — |
When to Choose AWS Bedrock
Cost-sensitive workloads
Use Amazon Nova Micro — AWS Bedrock's most affordable model at $0.04/M input / $0.14/M output. Estimated cost at 100K requests/month: $10.50.
Complex or high-accuracy tasks
Use Claude Sonnet 4.5 (Bedrock) — AWS Bedrock's most capable model (premium tier) with a 200,000-token context window. Priced at $3.00/M input / $15.00/M output.
High-volume pipelines
AWS Bedrock models are well-suited for high-throughput workloads. Consider batch API (where supported) for an additional 50% cost reduction on asynchronous pipelines. Check the pricing table above for batch pricing availability.
Compare AWS Bedrock Models vs Competitors
See detailed cost comparisons between AWS Bedrock models and models from other providers, including side-by-side pricing tables and monthly cost projections.
Use Cases for AWS Bedrock Models
Browse use case guides where AWS Bedrock models are recommended, including cost-effective model rankings and monthly cost estimates for each workload.
Frequently Asked Questions: AWS Bedrock API Pricing
What is the cheapest AWS Bedrock model?
The cheapest AWS Bedrock model by input token price is Amazon Nova Micro, priced at $0.04/M for input tokens and $0.14/M for output tokens. It is best suited for high-volume, cost-sensitive workloads where speed and price matter most.
What is the most capable AWS Bedrock model?
Claude Sonnet 4.5 (Bedrock) is AWS Bedrock's most capable model, classified as a premium-tier model. It supports a context window of 200,000 tokens and is priced at $3.00/M input / $15.00/M output per million tokens.
How does AWS Bedrock API pricing work?
AWS Bedrock charges per token consumed — separately for input tokens (the text you send) and output tokens (the text the model generates). Prices are listed in USD per 1 million tokens ($/M). AWS Bedrock was founded in 2023 and is headquartered in Seattle, WA. All pricing shown here is sourced from the official AWS Bedrock pricing page.
Does AWS Bedrock offer batch API pricing?
Yes — 8 AWS Bedrock models support batch processing at reduced rates: Claude Sonnet 4.5 (Bedrock), Claude 3.5 Sonnet v2 (Bedrock), Llama 3.1 70B (Bedrock), and others. Batch API is ideal for asynchronous workloads like bulk data extraction, document processing, or offline classification pipelines. Batch pricing is shown in the pricing table above.
Which AWS Bedrock models support vision / image inputs?
5 AWS Bedrock models support vision (image input): Claude Sonnet 4.5 (Bedrock), Claude 3.5 Sonnet v2 (Bedrock), Amazon Nova Pro, Amazon Nova Lite, Gemma 3 27B (Bedrock). These models can analyze images, charts, screenshots, and documents alongside text prompts.
AWS Bedrock pricing data last verified: 2026-03-11. Prices may change — verify on the official AWS Bedrock pricing page before production use.