Create and manage optimization jobs - Amazon Bedrock
Create a job (console)Create a job (Python SDK)Create a job (AWS CLI)Check job statusList jobsStop a jobBatch delete jobs
Services or capabilities described in AWS documentation might vary by Region. To see the differences applicable to the AWS European Sovereign Cloud Region, see the AWS European Sovereign Cloud User Guide.

Create and manage optimization jobs

Create a job (console)

To create a job using the console

  1. Give your advanced prompt optimization job a name.

  2. Select your target models (up to 5 per job). If you are migrating to a new model, select your current model as a baseline. Then select up to 4 other models you want to migrate to. If you are not changing models, just select your current model.

  3. Upload your JSONL input file from Amazon S3 or from your computer. Make sure you follow the Prepare your input dataset. Your input file can also include optimizationConfig to enable prompt shortening for a template. For more information, see Shorten prompts.

  4. For each prompt template, select your evaluation method. You can choose between steering criteria, LLM-as-a-judge, Lambda function, or leave blank for the system default.

  5. Select your output location in S3 for your raw results to be saved after the job.

  6. (Optional) Add a job description or a custom KMS key for encryption.

  7. Create the optimization.

Note

If you move the results files away from the output S3 location after the job is complete, the console results page will not render.

Create a job (Python SDK)

Replace the Region, S3 bucket location, model ID, and other fields as necessary for your use case. For modelId, you can specify a foundation model ID, a cross-region inference profile ID, or an application inference profile ARN.

import boto3
import json

client = boto3.client('bedrock', region_name='us-west-2')

response = client.create_advanced_prompt_optimization_job(
    jobName='my-optimization-job',
    modelConfigurations=[
        {'modelId': 'us.anthropic.claude-sonnet-4-5-20250929-v1:0'}
    ],
    inputConfig={
        's3Uri': 's3://my-bucket/input/dataset.jsonl'
    },
    outputConfig={
        's3Uri': 's3://my-bucket/output/'
    }
)

job_arn = response['jobArn']
print(f'Job created: {job_arn}')
Note

When you create the job through the API, no additional optimization configuration is needed in the CreateAdvancedPromptOptimizationJob request itself. Prompt shortening is determined per template from the input JSONL file. For more information, see Shorten prompts.

Create a job (AWS CLI)

Replace the Region, S3 bucket location, model ID, and other fields as necessary for your use case.

JOB_ARN=$(aws bedrock create-advanced-prompt-optimization-job \
    --job-name "my-optimization-job" \
    --model-configurations '[{"modelId": "us.anthropic.claude-sonnet-4-5-20250929-v1:0"}]' \
    --input-config '{"s3Uri": "s3://my-bucket/input/dataset.jsonl"}' \
    --output-config '{"s3Uri": "s3://my-bucket/output/"}' \
    --region us-west-2 \
    --query 'jobArn' --output text)
Note

When you create the job through the API, no additional optimization configuration is needed in the CreateAdvancedPromptOptimizationJob request itself. Prompt shortening is determined per template from the input JSONL file. For more information, see Shorten prompts.

Check job status

response = client.get_advanced_prompt_optimization_job(
    jobIdentifier=job_arn
)

print(f"Status: {response['jobStatus']}")
if response.get('failureMessage'):
    print(f"Failure: {response['failureMessage']}")
aws bedrock get-advanced-prompt-optimization-job \
    --job-identifier "$JOB_ARN" \
    --region us-west-2

List jobs

response = client.list_advanced_prompt_optimization_jobs(
    maxResults=10
)

for job in response['advancedPromptOptimizationJobSummaries']:
    print(f"{job['jobName']} | {job['jobStatus']} | {job['jobArn']}")
aws bedrock list-advanced-prompt-optimization-jobs \
    --max-results 10 \
    --region us-west-2

Stop a job

client.stop_advanced_prompt_optimization_job(
    jobIdentifier=job_arn
)
print('Job stop requested')
aws bedrock stop-advanced-prompt-optimization-job \
    --job-identifier "$JOB_ARN" \
    --region us-west-2

Batch delete jobs

response = client.batch_delete_advanced_prompt_optimization_jobs(
    jobIdentifiers=[job_arn]
)

for result in response['results']:
    print(f"{result['jobIdentifier']}: {result['status']}")
aws bedrock batch-delete-advanced-prompt-optimization-jobs \
    --job-identifiers "$JOB_ARN" \
    --region us-west-2