Claude Sonnet 5.5

Claude Sonnet 5.5 is Anthropic's September 28, 2026 Sonnet release for bounded coding and agent work, with a 1M context window and 128K maximum output.

Language ModelCoding AssistantAgent WorkflowsAdaptive ThinkingLong ContextAnthropic

Claude API / Amazon Bedrock / Google Cloud / Microsoft Foundry / Claude Platform on AWS

$2/1M input, $10/1M output, $0.20/1M cache reads; 5m cache writes $2.50 and 1h cache writes $4

Commercial

🚀Function Overview

Claude Sonnet 5.5 keeps a 1M context and 128K output limit, supports text and image input with text output, and uses adaptive thinking with high default effort.

Key Features

Official Claude API model ID claude-sonnet-5-5
1M-token context window and 128K maximum output
Text and image input with text output and tool use
Adaptive thinking with high default effort and between_tools for no up-front thinking
$2 input and $10 output per million tokens; $0.20 cache reads
$2.50 five-minute and $4 one-hour cache writes per million tokens
Available on Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS according to Anthropic
Migration from Sonnet 5 needs checks for thinking, forced tools, computer use, advisor models, and streaming behavior

Use Cases

•Well-scoped coding fixes
•Agent workflows with acceptance tests
•Document, slide, and spreadsheet drafting
•Vision-assisted analysis
•Sonnet 5 migration trials

⚙️Input Parameters

model

string

Use claude-sonnet-5-5 on the Claude API.

messages

array

Anthropic Messages API conversation payload with text or supported images.

tools

array

Optional supported tool definitions; forced tool use from Sonnet 5 must be migrated.

max_tokens

integer

Required maximum generated-token budget, up to the supported model limit.

output_config

object

Optional object with effort, for example {"effort":"high"}.

thinking

object

Optional adaptive thinking configuration, for example {"type":"adaptive"}.

💡Usage Examples

Example 1

Input Parameters

{
  "model": "claude-sonnet-5-5",
  "messages": [
    {
      "role": "user",
      "content": "Fix this bounded bug and report the checks you ran."
    }
  ],
  "max_tokens": 1024,
  "thinking": {
    "type": "adaptive"
  },
  "output_config": {
    "effort": "high"
  }
}

Output Results

Claude Sonnet 5.5 returns content blocks and tool calls. Verify the fix, latency, token use, cache behavior, and total accepted-task cost before changing production routing.

Quick Actions

Technical Specifications

Hardware Type: Claude API / Amazon Bedrock / Google Cloud / Microsoft Foundry / Claude Platform on AWS
Commercial Use: Supported
Pricing: $2/1M input, $10/1M output, $0.20/1M cache reads; 5m cache writes $2.50 and 1h cache writes $4