OpenAI models in Amazon Bedrock

Amazon Bedrock makes supported OpenAI models available through AWS-managed infrastructure. This deployment path is useful when your organization wants to keep procurement, identity, regional controls, and related cloud operations in AWS.

How Bedrock availability works

OpenAI models in Amazon Bedrock run through an AWS-managed deployment path with Responses API compatibility for supported models and capabilities. Your application still uses OpenAI model behavior, but AWS owns the surrounding cloud control plane, including account access, regional availability, and billing.

Use Bedrock when you need:

• AWS-native procurement and billing.
• AWS-managed identity, access, and account controls.
• Deployment in supported AWS Regions for customers with cloud-location requirements.

Use the OpenAI API directly when you need the broadest feature coverage, the latest first-party platform capabilities, or functionality unavailable in Bedrock.

Make Responses API requests

To send OpenAI SDK requests through Amazon Bedrock, use the Bedrock-aware SDK client and select the AWS Region and model ID for your deployment:

• Instantiate BedrockOpenAI instead of the default OpenAI client. The client derives the regional Mantle base URL from the AWS Region.
• For the initial openai.gpt-5.5 deployment, use us-east-2. This resolves to https://bedrock-mantle.us-east-2.api.aws/openai/v1.
• Use a Bedrock model ID with the openai. prefix, such as openai.gpt-5.5.

This example uses openai.gpt-5.5 in us-east-2. Use a supported model and AWS Region combination for your Bedrock deployment.

The following example uses a Bedrock API key stored as AWS_BEARER_TOKEN_BEDROCK. See Amazon Bedrock API keys for information about generating and using a Bedrock API key. The SDK reads the token from your environment.

1
2
3
4
5
6
7
8
9
10
11
12
import { BedrockOpenAI } from "openai";

const client = new BedrockOpenAI({
  awsRegion: "us-east-2",
});

const response = await client.responses.create({
  model: "openai.gpt-5.5",
  input: "Write a haiku about cloud infrastructure.",
});

console.log(response.output_text);

1
2
3
4
5
6
7
8
9
10
from openai import BedrockOpenAI

client = BedrockOpenAI(aws_region="us-east-2")

response = client.responses.create(
    model="openai.gpt-5.5",
    input="Write a haiku about cloud infrastructure.",
)

print(response.output_text)

1
2
3
4
5
6
7
curl "https://bedrock-mantle.us-east-2.api.aws/openai/v1/responses" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{
    "model": "openai.gpt-5.5",
    "input": "Write a haiku about cloud infrastructure."
  }'

For long-running applications, pass a token provider instead of a static API key. The SDK calls the provider before each request. The AWS token-generator packages return a cached short-term key when the current key is valid and generate a new key when needed. They use the AWS credential chain, which can include credentials configured with aws login.

Install the token-generator package for your SDK:

npm install @aws/bedrock-token-generator
pip install aws-bedrock-token-generator

1
2
3
4
5
6
7
8
9
10
11
12
13
14
import { getTokenProvider } from "@aws/bedrock-token-generator";
import { BedrockOpenAI } from "openai";

const client = new BedrockOpenAI({
  awsRegion: "us-east-2",
  bedrockTokenProvider: getTokenProvider(),
});

const response = await client.responses.create({
  model: "openai.gpt-5.5",
  input: "Write a haiku about cloud infrastructure.",
});

console.log(response.output_text);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
from aws_bedrock_token_generator import provide_token
from openai import BedrockOpenAI

client = BedrockOpenAI(
    aws_region="us-east-2",
    bedrock_token_provider=provide_token,
)

response = client.responses.create(
    model="openai.gpt-5.5",
    input="Write a haiku about cloud infrastructure.",
)

print(response.output_text)

Availability and operations

Availability depends on AWS Region and model. The initial launch scope is more limited than the OpenAI API, so check model support by AWS Region before rollout.

Amazon Bedrock provides Responses API-compatible inference for supported OpenAI models in supported AWS Regions in the United States. AWS manages authentication, account access, procurement, and billing.

AWS Regions are physical deployment locations, which differ from OpenAI data residency jurisdictions. Teams with residency requirements should evaluate the Bedrock Region itself and the corresponding AWS terms.

Responses API feature availability

Amazon Bedrock supports a subset of Responses API capabilities available through the OpenAI API. This table describes intended feature availability for the initial Amazon Bedrock offering. It excludes transient availability and service status.

Client-side tool_search is distinct from hosted tools and remote MCP server support. Hosted tools run through OpenAI-operated service infrastructure and are unavailable in the initial Amazon Bedrock offering.

Treat feature parity as workload-specific. If your application depends on a specific tool, response mode, or service tier, test that behavior through Bedrock before you commit to the deployment path.

Authentication and operations

Amazon Bedrock uses AWS-managed access controls. Your AWS administrator controls which accounts, roles, or temporary credentials can reach the supported model deployment. The exact authentication flow depends on the Bedrock configuration your organization uses.

Plan for AWS-owned operational checks such as:

• Account and model access configuration.
• Region-specific deployment approval.
• Temporary credential or token validity.
• AWS quota, logging, and support workflows.

Pricing

AWS bills Amazon Bedrock usage. Bedrock-specific pricing can differ from direct OpenAI API pricing, including regional processing premiums or other AWS-specific commercial terms.

See API pricing for direct OpenAI API pricing. For Bedrock pricing, use the AWS pricing materials published for the Bedrock deployment you plan to use.

Next steps

• Confirm your supported model and AWS Region in Amazon Bedrock.
• Verify the exact API features your workload needs.
• Compare Bedrock pricing and direct API pricing before launch.
• For Codex-specific setup, see Use Codex with Amazon Bedrock.