Claude Sonnet 5 Thinking

Claude Sonnet 5 with adaptive thinking enabled for complex coding, planning, and agentic work.

Reasoning
Vision
Native PDF input
Tool Calling
Structured Output

Added Jun 30, 2026

Pricing

Auto routing · per 1M tokens

Input: $2.00
Output: $10.00
Cache read: $0.20

Compare provider prices

Specifications

Context window: 1M
Max output: 128K
Avg output (7d): 1.2K tokens; Longer than 76% of models

Benchmarks

Artificial Analysis

LMArena

Sourced from Artificial Analysis.

Intelligence Index

38.2

Better than 91% of models compared

Coding Index

71.5

Better than 85% of models compared

Agentic Index

43.6

Better than 85% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

36.5%

Better than 55% of models compared

Harvey LAB-AA

Legal agentic work criterion pass rate

90.1%

Better than 56% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

1358 Elo

Better than 74% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

1465 Elo

Better than 82% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

13.2%

Better than 52% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

82.0%

Better than 92% of models compared

MLCR-AA

Medical long-context reasoning

55.0%

Better than 90% of models compared

Reasoning

HLE

Humanity's Last Exam

41.3%

Better than 89% of models compared

CritPt

Research-level physics reasoning

16.9%

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

14.1%

Better than 68% of models compared

SciCode

Python programming for scientific computing

54.3%

Better than 74% of models compared

Knowledge

AA-Omniscience Accuracy

Proportion of correctly answered questions

40.1%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

39.4%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

91.1%

Better than 91% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

82.0%

Better than 92% of models compared

GDPval-AA (unversioned / legacy)

Economically valuable tasks

48.3%

Last updated Oct 9, 2026