Claude Sonnet 5.5

Claude Sonnet 5.5 handles everyday coding, tool-based workflows, visual analysis, and polished writing with a 1M-token context window. This version skips up-front thinking and can reason between tool calls for faster responses; use Thinking for deeper planning and complex problems.

Reasoning
Vision
Native PDF input
Tool Calling
Structured Output

Added Sep 28, 2026

Pricing

Auto routing · per 1M tokens

Input: $2.00
Output: $10.00
Cache read: $0.10

Compare provider prices

Specifications

Context window: 1M
Max output: 128K
Avg output (7d): 1.1K tokens; Longer than 73% of models

Benchmarks

Artificial Analysis

LMArena

Design Arena

Sourced from Artificial Analysis.

Intelligence Index

46.8

Better than 96% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

59.4%

Better than 84% of models compared

AutomationBench-AA Tasks Completed

Fully completed workflows without guardrail violations

45.5%

Better than 93% of models compared

Harvey LAB-AA

Legal agentic work criterion pass rate

92.1%

Better than 73% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

1639 Elo

Better than 95% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

1551 Elo

Better than 87% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

25.2%

Better than 87% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

78.0%

Better than 79% of models compared

MLCR-AA

Medical long-context reasoning

75.0%

Better than 99% of models compared

Reasoning

HLE

Humanity's Last Exam

45.8%

Better than 93% of models compared

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

43.9%

Better than 89% of models compared

SciCode

Python programming for scientific computing

53.7%

Better than 69% of models compared

Legacy benchmarks

AA-LCR (unversioned / legacy)

Long context reasoning evaluation