Claude Sonnet 5.5
Amazon Bedrock token pricing and cost optimization reference for Claude Sonnet 5.5.Rates normalized to USD per 1 Million Tokens ($/1M) with prompt caching and batch API discounts across 33 AWS regions.
anthropic.claude-sonnet-5-5-v1:0Alternatives to Claude Sonnet 5.5
| Option | Price vs Claude Sonnet 5.5 | Why look at it |
|---|---|---|
| Claude Haiku 5.5 | No like-for-like unit | Also from the same provider |
| Claude Opus 5.5 | No like-for-like unit | Also from the same provider |
| Claude Sonnet 5 | No like-for-like unit | Also from the same provider |
What to know about Claude Sonnet 5.5
Anthropic's mid-priced Claude model on Bedrock: $2.00 input and $10.00 output per million tokens on the global endpoint.
Claude Sonnet 5.5 sits in the middle of Anthropic’s lineup on Amazon Bedrock. On the global endpoint in us-east-1 it costs $2.00 per million input tokens and $10.00 per million output tokens.
Where the traffic runs affects the price. In-geography inference costs $2.20 and $11.00, a 10% premium, and is the choice if data has to stay in a region. Use global if you can.
Output costs 5 times the input rate, so verbose responses drive the bill more than long prompts do. Prompt caching helps on the input side: cache reads cost $0.10 per million tokens, 5% of the normal input price. Writing to the cache costs $2.50 per million tokens, or $4.00 for a one-hour cache. Caching pays off when many requests share a long prefix, such as a system prompt or a document.
Opus 5.5 costs $4.00 and $20.00, twice as much. Haiku 5.5 costs $0.10 and $0.50, one twentieth. Send simple classification or extraction to Haiku and keep Sonnet for work that needs it.
Sonnet 5 has the same base token prices, but its cache reads cost $0.20, twice the 5.5 rate.
Good for
- General-purpose assistants and agent workloads
- Prompts that reuse a large prefix, using prompt caching
- A step down from Opus 5.5 at half the token price
Watch for
- In-geography endpoints add 10% over global
- Output tokens cost 5 times input tokens
- Haiku 5.5 is far cheaper if the task is simple
Claude Sonnet 5.5 Spend Estimator
Model inference spend by prompt volume, cache hit rate, and batch processing.
Key Cost Levers & Architecture
Practical mechanisms to reduce inference costs with Claude Sonnet 5.5 on AWSCaches repetitive system instructions, schemas, and document prefixes in memory with 5-minute and extended 1-hour TTL tiers.
- Read Rate: $0.10 / 1M tokens
- Write Rate (5m TTL): $2.50 / 1M tokens
- Write Rate (1h TTL): $4.00 / 1M tokens
- Ideal for: Multi-turn conversations, RAG document queries, structured schemas
Asynchronous bulk processing submitted via Amazon S3 with a standard 24-hour turnaround SLA.
- Batch Input: $1.00 / 1M tokens
- Batch Output: $5.00 / 1M tokens
- Ideal for: Offline classification, synthetic data generation, benchmark evals
AWS Bedrock supports standard on-demand inference, priority reserved throughput, and cross-region routing.
- Standard: Lowest latency, pay-per-token with zero commitments
- Cross-Region Routing: Automatically burst traffic to available regional capacity
- GovCloud: Dedicated isolated compliance partitions with +20% uplift
Regional Pricing Matrix
Live rates across all 33 available AWS cloud regions| AWS Region | Prompt Input $/1M | Completion Output $/1M | Cache Read $/1M | Batch Input $/1M | Pricing Tier |
|---|---|---|---|---|---|
| us-east-1 US East (N. Virginia) | $2.00 | $10.00 | $0.10 | — | Baseline |
| us-west-2 US West (Oregon) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-central-1 Europe (Frankfurt) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-west-1 Europe (Ireland) | $2.00 | $10.00 | $0.10 | — | Baseline |
| us-gov-west-1 AWS GovCloud (US-West) | $2.40 | $12.00 | $0.12 | — | GovCloud (+20%) |
| af-south-1 Africa (Cape Town) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-south-2 Asia Pacific (Hyderabad) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-3 Asia Pacific (Jakarta) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-5 Asia Pacific (Malaysia) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-4 Asia Pacific (Melbourne) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-south-1 Asia Pacific (Mumbai) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-6 Asia Pacific (New Zealand) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-northeast-2 Asia Pacific (Seoul) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-1 Asia Pacific (Singapore) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-2 Asia Pacific (Sydney) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-east-2 Asia Pacific (Taipei) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ap-southeast-7 Asia Pacific (Thailand) | $2.00 | $10.00 | $0.10 | — | Baseline |
| us-gov-east-1 AWS GovCloud (US-East) | $2.40 | $12.00 | $0.12 | — | GovCloud (+20%) |
| ca-central-1 Canada (Central) | $2.00 | $10.00 | $0.10 | — | Baseline |
| ca-west-1 Canada West (Calgary) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-west-2 Europe (London) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-south-1 Europe (Milan) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-west-3 Europe (Paris) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-south-2 Europe (Spain) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-north-1 Europe (Stockholm) | $2.00 | $10.00 | $0.10 | — | Baseline |
| eu-central-2 Europe (Zurich) | $2.00 | $10.00 | $0.10 | — | Baseline |
| il-central-1 Israel (Tel Aviv) | $2.00 | $10.00 | $0.10 | — | Baseline |
| mx-central-1 Mexico (Central) | $2.00 | $10.00 | $0.10 | — | Baseline |
| me-south-1 Middle East (Bahrain) | $2.00 | $10.00 | $0.10 | — | Baseline |
| me-central-1 Middle East (UAE) | $2.00 | $10.00 | $0.10 | — | Baseline |
| sa-east-1 South America (Sao Paulo) | $2.00 | $10.00 | $0.10 | — | Baseline |
| us-east-2 US East (Ohio) | $2.00 | $10.00 | $0.10 | — | Baseline |
| us-west-1 US West (N. California) | $2.00 | $10.00 | $0.10 | — | Baseline |
Developer Quickstart
Invoke Claude Sonnet 5.5 via Amazon Bedrock Converse APIimport boto3
# Amazon Bedrock Converse API invocation
client = boto3.client("bedrock-runtime", region_name="us-east-1")
response = client.converse(
modelId="anthropic.claude-sonnet-5-5-v1:0",
messages=[
{
"role": "user",
"content": [{"text": "Summarize key features and cost levers of Claude Sonnet 5.5."}]
}
],
inferenceConfig={
"maxTokens": 1024,
"temperature": 0.7
}
)
output_text = response["output"]["message"]["content"][0]["text"]
usage = response["usage"]
print(f"Response: {output_text}")
print(f"Usage: {usage['inputTokens']} in, {usage['outputTokens']} out")