[FEATURE] Add support for AWS Bedrock service_tier parameter (flex, priority, reserved, default)

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

Support AWS Bedrock Service Tier Selection
https://docs.aws.amazon.com/bedrock/latest/userguide/service-tiers-inference.html

### Summary
Add a configuration option to set the AWS Bedrock service_tier parameter when using Claude CLI with Bedrock, allowing users to leverage Bedrock's different pricing and performance tiers (flex, priority, default, reserved).

### Motivation
AWS Bedrock supports different service tiers with varying pricing and latency characteristics:

• Flex: Cost-effective tier with pricing discounts for workloads that tolerate longer processing times
• Priority: Fastest response times at premium pricing
• Default: Standard tier with consistent performance
• Reserved: Fixed-price capacity reservations for mission-critical workloads

Currently, Claude CLI does not expose the service_tier parameter, forcing users to use the default tier or bypass Claude CLI entirely to use cost-effective options like flex tier.

Proposed Solution

Option 1: Environment Variable
export CLAUDE_CODE_USE_BEDROCK=1
export AWS_BEDROCK_SERVICE_TIER=flex # or priority, default, reserved

Option 2: Command-line Flag
claude --bedrock-service-tier flex

Option 3: Settings File
{
"bedrock": {
"serviceTier": "flex"
}
}

Technical Details

The service_tier parameter is sent as part of the Bedrock runtime API request:

{
"service_tier": "flex",
"anthropic_version": "bedrock-2023-05-31",
"messages": [...]
}

Reference: https://docs.aws.amazon.com/bedrock/latest/userguide/service-tiers-inference.html

Alternative Solutions

_No response_

Priority

High - Significant impact on productivity

Feature Category

API and model interactions

Use Case Example

1. Cost Optimization: Use flex tier for batch processing, evaluations, and non-time-sensitive tasks
2. Performance Requirements: Use priority tier for latency-sensitive applications
3. Budget Control: Leverage reserved capacity for predictable workloads

Additional Context

Service tier is an optional parameter (defaults to standard tier if not specified)