Anthropic

Claude Sonnet 5.5

Sonnet for everyday coding and agents. Stronger results than Sonnet 5 at the same token price.

successor to claude-sonnet-5text + image inputtext output

Replaces Claude Sonnet 5: same input and output price, same context window.

Run this model

anthropic/claude-sonnet-5.5

Use an endpoint supported by this model.

Get a key for this setup →
const response = await fetch('https://api.minirouter.sh/v1/chat/completions', {  method: 'POST',  headers: {    Authorization: `Bearer ${process.env.MINIROUTER_KEY}`,    'Content-Type': 'application/json',  },  body: JSON.stringify({    model: 'anthropic/claude-sonnet-5.5',    messages: [{ role: 'user', content: 'Why is the sky blue?' }],    max_tokens: 1024,  }),}) const data = await response.json()console.log(data.choices[0].message.content)
Read docs →

Use text or image input; PDF requests are not supported. Remove temperature and top_p when migrating. See the migration notes for native thinking and tool settings.

Model overview

Input / 1M
$2.1
All-in MiniRouter rate
Output / 1M
$10.5
All-in MiniRouter rate
Context
1M
1,000,000 tokens
Max output
128K
tokens
Terminal-Bench 4.0
70.6%
Anthropic-reported
Effort
low → max
five settings
Announced
28 Sept 2026
publisher release

Publisher-reported, plotted by MiniRouter

A stronger Sonnet for coding.

Compare Anthropic's launch scores. MiniRouter did not run these evaluations.

Agentic coding

Terminal-Bench 4.0

Opus 5.5 uses xhigh, its best reported effort setting.

Published score

0 to 100%

  • Claude Sonnet 5.570.6%
  • Claude Sonnet 510.3%
  • Claude Opus 5.566.4%
Source: Anthropic announcement ↗, September 28, 2026.
View chart data
Terminal-Bench 4.0 published values
ModelScore
Claude Sonnet 5.570.6%
Claude Sonnet 510.3%
Claude Opus 5.566.4%

Agentic coding

CursorBench 4.0

Tasks drawn from real Cursor coding sessions.

Published score

0 to 100%

  • Claude Sonnet 5.555.5%
  • Claude Sonnet 534.1%
  • Claude Opus 5.557.8%
Source: Anthropic announcement ↗, September 28, 2026.
View chart data
CursorBench 4.0 published values
ModelScore
Claude Sonnet 5.555.5%
Claude Sonnet 534.1%
Claude Opus 5.557.8%

Agentic coding

FrontierCode 1.1 (Main)

Sonnet 5.5 at xhigh. Max scored 46.2%; extra work can exceed task scope.

Published score

0 to 100%

  • Claude Sonnet 5.552.1%
  • Claude Sonnet 542.4%
  • Claude Opus 5.554.4%
Source: Anthropic announcement ↗, September 28, 2026.
View chart data
FrontierCode 1.1 (Main) published values
ModelScore
Claude Sonnet 5.552.1%
Claude Sonnet 542.4%
Claude Opus 5.554.4%

Publisher list prices

Same rates. Fewer tokens per task.

Anthropic reports up to 30% lower task cost and 30%+ faster output than Sonnet 5. Your workload and route determine the result.

Anthropic list, per 1M tokens

Input
$2.00
Output
$10.00
Cache read
$0.20
5m cache write
$2.50
Source: Anthropic model pricing ↗. Current MiniRouter rates are shown above.
View chart data
Anthropic list prices per million tokens
RateSonnet 5.5Opus 5.5
Input$2.00$4.00
Output$10.00$20.00
Cache read$0.20$0.20
5m cache write$2.50$5.00

Choose and configure

Match effort to the work.

Start at medium for defined coding tasks. Anthropic still recommends Opus 5.5 for open-ended work requiring sustained judgment.

Anthropic migration guide
  1. 01

    Review request options

    The native Claude API rejects forced tool choice, thinking budgets and non-default sampling settings. Use automatic tool choice and adaptive thinking.

  2. 02

    Replace disabled thinking

    The native between_tools setting replaces disabled, at high effort or below. Higher effort requires adaptive thinking.

  3. 03

    Preserve tool history

    Keep signed thinking blocks unchanged and conversations append-only. Recheck token budgets: thinking counts toward output usage.

One route, standard clients.

Authenticate with a MiniRouter bearer key. Caller-provided upstream credentials are not used on this route.

Read model setup docs →
Base URLhttps://api.minirouter.sh/v1
Model IDanthropic/claude-sonnet-5.5
Endpoints
/v1/chat/completions/v1/messages/v1/responses
Parameters

max_tokens, stop, tools, tool_choice, reasoning, include_reasoning

Structured output

Not enforced; a requested format is dropped and the answer is plain text.

Configured provider

Priority follows the live catalog order.

Input: text, image · Output: text

ProviderPriorityInput / 1MOutput / 1MParametersAccess
vercel
1$2.1$10.5max_tokens, stop, tools, tool_choice, reasoning, include_reasoningMiniRouter balance