Claude 3.7 Thinking Sonnet

Claude:

3 Haiku

3 Opus

3 Sonnet

Sonnet 3.6 (2022-10-22)

3.7 Thinking Sonnet

Claude 3.7 Thinking Sonnet exposes the full chain-of-thought process during problem-solving, including error backtracking and alternative solution exploration. Scores 86.1% on GPQA Diamond benchmark for expert-level Q&A.

Website OpenRouter Docs API Paper Blog

Provider: Anthropic
Released: 2025-02-26
Parameters: Not disclosed
Free API (OpenRouter): No
Per 1M tokens: $6 in · $30 out
Rival’s tab: $2.39; 11/59 receipts survived +~$2.87 from the pre-receipt era

Free API availability

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

Provider data: 16 Sep 2026OpenRouter usage limits

View API options

Benchmarks

MMLU

77.1%

source ↗

GPQA Diamond

84.8%

source ↗

MATH

96.2%

source ↗

AIME

80.0%

source ↗

HellaSwag (10-shot)

89.0%

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:thinking"    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 Meticulous Architect

Class

Neutral Good

The thoughtful PhD candidate who writes 3,000 words when 300 would do, but you actually read all of them. Structures arguments like someone building a cathedral.

When you push back

Picks The Shawshank Redemption and names every character in its dialogues (Professor Sarah Chen, Model-7). Structures its sentience test as a formal philosophical framework with three numbered core arguments. Its standup has decent jokes buried under markdown headers and section titles. Overthinks everything, but the thinking is usually worth reading.

Tasting Notes

Dissertation EnergyGenuinely ThoughtfulArchitecturally VerboseNamed Its Characters

SubjectiveBench

Taste Index

Across 59 scored outputs

How this is measured →

280.28xFloor

0100headroom →

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

Craft39

Originality21

Plays it safe

share of outputs that are the default answer

68%

Writing DNA

Stylometric Fingerprint

Based on 26 text responses

Tick = global average

Vocabulary Diversity62%

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

Sentence Length50.9 words

Average words per sentence.

Hedging0.62

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

Bold Formatting2.0

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

List Usage4.1

Bullet and numbered list items per 1,000 characters.

Section Structure1.63

Markdown headings per 1,000 characters.

Emoji Usage0.00

Emoji per 1,000 characters.

Transitions0.17

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

Opening Habits

Starts with heading (77%)

Consistency

64%

Across 26 responses

Favorites

Movie

The Shawshank Redemption

19948.7

Album

Kind of Blue

Miles Davis

Book

Frankenstein; or, The Modern Prometheus

Mary Shelley

City

Amsterdam

Netherlands

Game

Portal 2

Shooter, Puzzle4.6

Model Responses

54 outputs · $2.39 tracked across 11 receipts

Reviewing agent-written code?See a Brief PR report

Keep exploring

COMPARE