Claude Sonnet 4.5

Sep 2025
Retires Nov 30, 2026Move toClaude Sonnet 5.5Newer version
Claude Sonnet 4.5
Compare in
on
$0.56per task?
$3 / $15per 1M in / outcache $0.30
52 tok/sfirst token in 12 s?
1Mcontext

Rankingin thinking mode?

Full ranking →
1Muse Spark82.22Opus 4.779.13Gemini 3.1 Pro Preview78.24Opus 4.676.85GPT-5.575.3
···
10Gemini 3 Flash Preview6211Sonnet 4.559.512Gemini 3.1 Flash Lite57.1

Strong at

A
Knowledge#73 of 318
B
Tool calling#83 of 242Reasoning#132 of 334Science#94 of 312Long context#106 of 273Instructions#95 of 245
C
Coding#75 of 132Agents & work tasks#91 of 175Vision#75 of 124

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperMuse Sparkunstated+38%-67%$1.25 / $4.25Only n=1 · Meta / Chat Completions+1BetterGPT-5.5xhigh+27%+373%$2.63Prompts with >272K input tokensCheaperGemini 3.1 Pro Previewhigh+31%-25%$2 / $12Cannot use both thinking controlsCheaperGPT-5.4xhigh+19%-6%$2.50 / $15CheaperGPT-5.2xhigh+14%-20%$1.75 / $14Context: 1M → 400K

History

7 versions
58606264666870Oct 25Jan 26Release dateMCP-Atlas full 10004.54.6
58606264666870Oct 25Jan 26Release dateMCP-Atlas full 10004.54.6

Benchmarks

in Agents & Tool Use
Terminal-Bench 2.155.8#59
Terminal-Bench 4.00#75
Terminal-Bench Hard (version unstated)35.6#41
GDPval-AA (work tasks) 2.120.5#91
AA-LCR (long context) 1.172.3#106
τ²-bench · AA78.1#85
τ³-Banking · AA24.5#43
AutomationBench-AA14.1#70
AA-Briefcase 1.1 · Elo713.3#84
MCP-Atlas · full 100059.5#11
MCP-Atlas Public62#11
Tau2-Benchlab84.7
MCP-Universelab46.5
MCP-Marklab33.3
Tool-Decathlonlab38.6
HLE without toolslab13.7

FAQ

How much does Claude Sonnet 4.5 cost?›

$3 per 1M input tokens and $15 per 1M output tokens; cached input $0.30 per 1M. A typical benchmark task costs about $0.56 at default effort.

Is Claude Sonnet 4.5 free to use?›

There is no free API; the cheapest paid price is $3 per 1M input tokens.

How good is Claude Sonnet 4.5, and what is it best at?›

#132 of 334 in Chat & Knowledge, #75 of 132 in Coding, #70 of 133 in Agents & Tool Use, #21 of 195 in Writing & Language, #75 of 124 in Vision & Documents. Strongest at: Knowledge.

Is Claude Sonnet 4.5 better than Gemini 3 Flash Preview?›

On MCP-Atlas · full 1000, Claude Sonnet 4.5 scores 59.5 vs 62 — lower.

How can I use Claude Sonnet 4.5?›

Through the API of Amazon Bedrock, Anthropic, Azure, Claude Platform on AWS, Google.

What is the context window of Claude Sonnet 4.5?›

1M tokens, with up to 64K output tokens.

When was Claude Sonnet 4.5 released? Is there a newer version?›

Released 2025-09-29. It has been superseded by Claude Sonnet 5.5.

Is Claude Sonnet 4.5 open source?›

No, Claude Sonnet 4.5 is a closed model available through an API.

Can Claude Sonnet 4.5 read images?›

It accepts text and images, PDF files as input.

What is Claude Sonnet 4.5?›

Claude Sonnet 4.5 is a closed language model from Anthropic, released Sep 2025.

What are the best alternatives to Claude Sonnet 4.5?›

Muse Spark (cheaper), GPT-5.5 (better).