Does boto3 get_object(...) immediately download the actual file in python?

Background:

According to the boto3 get_object() documentation, the method get_object(**kwargs) has the following response signature:

{
    'Body': StreamingBody(),
    'DeleteMarker': True|False,
    'AcceptRanges': 'string',
    'Expiration': 'string',
    'Restore': 'string',
    'LastModified': datetime(2015, 1, 1),
    'ContentLength': 123,
    'ETag': 'string',
    'MissingMeta': 123,
    'VersionId': 'string',
    'CacheControl': 'string',
    'ContentDisposition': 'string',
    'ContentEncoding': 'string',
    'ContentLanguage': 'string',
    'ContentRange': 'string',
    'ContentType': 'string',
    'Expires': datetime(2015, 1, 1),
    'WebsiteRedirectLocation': 'string',
    'ServerSideEncryption': 'AES256'|'aws:kms',
    'Metadata': {
        'string': 'string'
    },
    'SSECustomerAlgorithm': 'string',
    'SSECustomerKeyMD5': 'string',
    'SSEKMSKeyId': 'string',
    'StorageClass': 'STANDARD'|'REDUCED_REDUNDANCY'|'STANDARD_IA'|'ONEZONE_IA'|'INTELLIGENT_TIERING'|'GLACIER'|'DEEP_ARCHIVE',
    'RequestCharged': 'requester',
    'ReplicationStatus': 'COMPLETE'|'PENDING'|'FAILED'|'REPLICA',
    'PartsCount': 123,
    'TagCount': 123,
    'ObjectLockMode': 'GOVERNANCE'|'COMPLIANCE',
    'ObjectLockRetainUntilDate': datetime(2015, 1, 1),
    'ObjectLockLegalHoldStatus': 'ON'|'OFF'
}

Question:

Currently, I just need the data contained in the Metadata field. But I want to make sure I am not incurring the unnecessary cost of downloading the contents of Body.

Does boto3 immediately download the contents of Body or does it wait until I call response["Body"].read()?

If you only want the object's metadata, use head_object():

The HEAD operation retrieves metadata from an object without returning the object itself. This operation is useful if you're only interested in an object's metadata. To use HEAD, you must have READ access to the object.