Anthropic logo

Claude Sonnet 5

Anthropic · Released Jun 30, 2026

ProprietarycodingagentsreasoningCompare this model

$2

Input price / MTok

#17 of 27

$10

Output price / MTok

#19 of 27

1.0M

Context window

#5 of 28

79

Tokens / sec

#15 of 21

63.2

SWE-bench Pro

#3 of 9

81.2

OSWorld

#3 of 8

Summary

Claude Sonnet 5 is a proprietary language model from Anthropic, released on Jun 30, 2026.

At $2 input and $10 output per million tokens, it ranks #17 of 27 for input price, placing it mid-range on input cost.

Its context window of 1.0M tokens ranks #5 of 28 for context depth in this dataset.

It targets coding, agents, reasoning workloads.

Throughput is measured at 79 tokens per second (ranked #15 of 21 for throughput).

Default model for Claude Free and Pro plans; adaptive thinking always on.

Benchmarks

BenchmarkScoreMax%RankSource
SWE-bench Pro63.210063.2%#3/9Anthropic (self-reported in launch annou...
OSWorld81.210081.2%#3/8Anthropic (self-reported in launch annou...

Similar Models

Compare all 6 models

Models with the most shared benchmark coverage and closest scores to Claude Sonnet 5.

Claude Sonnet 5
OpenAIGPT-5.5
AnthropicClaude Opus 4.8
GoogleGemini 3.5 Flash
OpenAIGPT-5.4
AnthropicClaude Sonnet 4.6
SWE-bench Pro#3 of 9
Professional software engineering tasks
OSWorld#3 of 8
Computer-use and GUI automation

Compare Claude Sonnet 5 side by side

Pre-filled with 5 similar models

Open comparison

Speed source: https://artificialanalysis.ai/models/claude-sonnet-5