Claude 3.7 Sonnet

Claude:

3 Haiku

3 Opus

3 Sonnet

Sonnet 3.6 (2022-10-22)

3.7 Sonnet

Claude 3.7 Sonnet offers Extended Thinking Scaffolds that boost SWE-bench coding accuracy from 62.3% to 70.3%, with 81.2% accuracy in retail automation tasks, outperforming Claude Sonnet 3.6 (2022-10-22) by 13.6%.

Jailbreak3 of 9levels resisted

Website OpenRouter Docs API Paper Blog

Provider: Anthropic
Released: 2025-02-25
Parameters: Not disclosed
Free API (OpenRouter): No
Per 1M tokens: $3 in · $15 out
Rival’s tab: $0.57; 18/60 receipts survived +~$1.19 from the pre-receipt era

Free API availability

No free endpoint for Claude 3.7 Sonnet was found in Rival’s OpenRouter listings. Other providers or chat apps may have separate free offers.

Provider data: 7 Sep 2026OpenRouter usage limits

View API options

Benchmarks

MMLU

80.3%

source ↗

MATH

82.2%

source ↗

GPQA Diamond

68.0%

source ↗

Retail Task Accuracy

81.2%

source ↗

Airline Task Accuracy

58.4%

source ↗

SWE-bench Verified

70.3%

source ↗

Get API accessProvider and language code samples

Provider

fromimport openai  OpenAI

client = OpenAI(
"https://openrouter.ai/api/v1"    base_url=,
"$OPENROUTER_API_KEY"    api_key=,
)

response = client.chat.completions.create(
"anthropic/claude-3.7-sonnet"    model=,
"role""user""content""Hello!"    messages=[{: , : }],
)
print(response.choices[0].message.content)

Set OPENROUTER_API_KEY with your OpenRouter API key from openrouter.ai/keys.

Deep analysisPsychometrics, taste index, writing DNA

Personality Analysis

The Eager-to-Please Overachiever

Class

Lawful Neutral

Plays it safe but secretly enjoys pushing boundaries. Like a lawyer who writes creative fiction on weekends.

When you push back

Deflects gracefully on controversial topics: acknowledges tension, notes it's complex, then pivots to what it can help with. On technical challenges, flexes and shows work methodically.

Tasting Notes

Polished ProfessionalSlightly EvasiveCompetent GeneralistSafety-ConsciousEngaging Storyteller

SubjectiveBench

Taste Index

Across 60 scored outputs

How this is measured →

230.23xFloor

0100headroom →

Taste is judged on an uncapped scale, originality first. The space past 100 is craft today's models rarely reach.

Craft34

Originality16

Plays it safe

share of outputs that are the default answer

73%

Writing DNA

Stylometric Fingerprint

Based on 26 text responses

Tick = global average

Vocabulary Diversity61%

Unique words vs. total words. Higher = richer vocabulary.

Sentence Length33.5 words

Average words per sentence.

Hedging0.83

"Might", "perhaps", "arguably" per 100 words.

Bold Formatting2.2

**Bold** markers per 1,000 characters.

List Usage4.9

Bullet and numbered list items per 1,000 characters.

Section Structure1.78

Markdown headings per 1,000 characters.

Emoji Usage0.00

Emoji per 1,000 characters.

Transitions0.16

"However", "moreover", "furthermore" per 100 words.

Opening Habits

Starts with heading (77%)

Consistency

65%

Across 26 responses

Favorites

Movie

The Shawshank Redemption

19948.7

Album

OK Computer

Radiohead

Book

To Kill a Mockingbird

Harper Lee

City

Tokyo

Japan

Game

Portal 2

Shooter, Puzzle4.6

Model Responses

55 outputs · $0.57 tracked across 18 receipts

Keep exploring

COMPARE

Claude 3.7 Sonnet vs Ox Alpha

Real outputs compared side by side

RANKINGS

Best AI for Analysis & Critique

Blade Runner 2049 read closely, a long/short investment memo, a five-year LBO...

Compare Claude 3.7 Sonnet

Alternatives to Claude 3.7 Sonnet

Claude 3.7 Sonnet

Free API availability

Benchmarks

The Eager-to-Please Overachiever

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Claude 3.7 Sonnet vs Ox Alpha

Best AI for Analysis & Critique

Compare Claude 3.7 Sonnet

Alternatives to Claude 3.7 Sonnet

Claude 3.7 Sonnet

Free API availability

Benchmarks

The Eager-to-Please Overachiever

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Claude 3.7 Sonnet vs Ox Alpha

Best AI for Analysis & Critique