Skip to content
ModelCap

Inference provider

Amazon Bedrock: AI models and API pricing

Every tracked model Amazon Bedrock serves through OpenRouter, with Amazon Bedrock's own input and output price per 1M tokens, the context window it serves, disclosed quantization and measured endpoint uptime. Rows link to the model page, where every provider's deployment is listed side by side.

Snapshot as of 11 August 2026

Amazon Bedrock serves 28 tracked models through 45 endpoints as of 11 August 2026; the highest-ranked is Claude Fable 5 (ModelCap Index #1); mean measured uptime is 96.65%.

Models served
28
On the ModelCap Index
14
Cheapest output price
$0.14/1M

Language models on Amazon Bedrock (28)

Ranked models first, in ModelCap Index order. Prices are the lowest Amazon Bedrock lists for the model; the endpoint count shows how many deployments carry it.

Language models served by Amazon Bedrock, with price, context, quantization and uptime
#ModelModelCap IndexInput $/1MOutput $/1MContextQuantizationUptime (30 min)Endpoints
1Claude Fable 5Anthropic#183.1$10.00$50.001M—98.95%2
2Claude Opus 5Anthropic#279.6$5.00$25.001M—99.66%3
3GPT-5.6 SolOpenAI#777.3$5.50$33.001M——1
4GPT-5.5OpenAI#1073.9$5.50$33.001M——1
5Claude Sonnet 5Anthropic#1470.0$2.00$10.001M—100.0%3
6GPT-5.6 TerraOpenAI#1569.5$2.20$13.201M——1
7GPT-5.6 LunaOpenAI#1963.3$0.22$1.321M——1
8Claude Haiku 4.5Anthropic#3651.0$1.00$5.00200K—99.87%2
9Qwen3 Coder 30B A3B InstructQwen#7428.4$0.15$0.60———1Open weights
10Nova 2 LiteAmazon#8027.1$0.30$2.501M—99.93%2
11Palmyra X5Writer#17225.5$0.60$6.001M——1
12Nova Pro 1.0Amazon#19914.8$0.80$3.20300K——2
13Nova Micro 1.0Amazon#2137.3$0.035$0.14128K—61.57%2
14Nova Premier 1.0Amazon#2300.0$2.50$12.501M——1
15Claude 3 HaikuAnthropic—$0.25$1.25200K—100.0%1
16Claude Opus 4.1Anthropic—$15.00$75.00200K—100.0%1
17Claude Opus 4.5Anthropic—$5.00$25.00200K—100.0%2
18Claude Opus 4.6Anthropic—$5.00$25.001M—100.0%1
19Claude Opus 4.7Anthropic—$5.00$25.001M—100.0%2
20Claude Opus 4.8Anthropic—$5.00$25.001M—100.0%2
21Claude Sonnet 4Anthropic—$3.00$15.00200K—99.94%2
22Claude Sonnet 4.5Anthropic—$3.00$15.001M—99.94%3
23Claude Sonnet 4.6Anthropic—$3.00$15.001M—99.43%2
24GLM 5Z.ai—$1.00$3.20203K—99.66%1Open weights
25GPT-5.4OpenAI—$2.75$16.501M——1
26Kimi K2.5Moonshot AI—$0.60$3.00262K—99.69%1
27Llama 3.1 70B InstructMeta—$0.72$0.72131K—100.0%1
28Nova Lite 1.0Amazon—$0.06$0.24300K—100.0%2

Amazon Bedrock on ModelCap: common questions

How many AI models does Amazon Bedrock serve?

28 tracked models as of 11 August 2026: 28 language models (14 with a public ModelCap Index position), across 45 live endpoints observed through OpenRouter.

What is the best model available on Amazon Bedrock?

Claude Fable 5 from Anthropic is the highest-ranked language model Amazon Bedrock serves, at ModelCap Index position #1 and $50.00 per 1M output tokens on Amazon Bedrock. The table lists every model in Index order.

What is the cheapest model on Amazon Bedrock?

Nova Micro 1.0 at $0.14 per 1M output tokens ($0.035 per 1M input), the lowest positive language-token price Amazon Bedrock lists. Free-tier endpoints, where present, show as $0 and are excluded from that comparison.

Which Amazon Bedrock model has the largest context window?

GPT-5.6 Sol, served with a 1M-token context window on Amazon Bedrock. Providers sometimes serve less than the model's published maximum, so the figure here is the provider's own.

How reliable is Amazon Bedrock?

Mean measured uptime across Amazon Bedrock endpoints is 96.65% over the last 30 minutes, as reported by OpenRouter at the time of the snapshot. Per-model uptime is in the table; it is a live operational signal, not a service-level guarantee.

Where do the Amazon Bedrock prices come from?

From the OpenRouter catalogue entry for each Amazon Bedrock endpoint, in USD per 1M tokens, re-observed on every ModelCap refresh. Where Amazon Bedrock exposes several deployments of one model at different prices, the table shows the lowest and notes the count; the model page lists each deployment.

How current is this page?

It re-renders within a minute of every dataset refresh; this snapshot was published 11 August 2026. Offers retained from a failed refresh are marked stale in the data and are not counted here.

Explore further