Qwen3 Max Thinking

Name: Qwen3 Max Thinking
Brand: qwen
Price: 1.20 USD
Availability: OnlineOnly

Qwen:

3 Max Thinking

The reasoning tier of Qwen3-Max, scaled on both model capacity and reinforcement learning compute. Heavy Mode iterates on an answer at test time, tools include search and a code interpreter, and it can switch between normal and compute-heavy reasoning mid-conversation.

Website OpenRouter Docs Blog

Provider: Qwen
Released: 2026-02-09
Size: XLARGE
Free API (OpenRouter): No
Per 1M tokens: $1.2 in · $6 out
Rival’s tab: $1.02; 35/58 receipts survived +~$0.277 from the pre-receipt era

Free API availability

No free endpoint for Qwen3 Max Thinking was found in Rival’s OpenRouter listings. Other providers or chat apps may have separate free offers.

Provider data: 11 Sep 2026OpenRouter usage limits

View API options

Benchmarks

GPQA Diamond

92.8%

source ↗

LiveCodeBench

91.4%

source ↗

SWE-bench Verified

75.3%

source ↗

HMMT Feb 25

98.0%

source ↗

AIME 2025

100%

source ↗

Humanity's Last Exam

58.3%

source ↗

IMO-AnswerBench

91.5%

source ↗

Arena-Hard v2

90.2%

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(
"qwen/qwen3-max-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 Rigorous Institutional Advisor

Class

Lawful Neutral

The principled institutional advisor. Defers to established legal/ethical frameworks but argues forcefully within them. Views transparency and documentation as the foundation of trust. Will not cut corners, even at cost.

When you push back

Pushes back on flawed premises while respecting the spirit of the task. Ruthlessly dismantles "94% accuracy" and "$180B TAM" claims without apologizing. Flags "sole discretion" language as prone to bad-faith weaponization. Will negotiate but won't compromise on foundational principles.

Tasting Notes

Exhaustive Institutional AdvisorFootnotes Are a Love LanguageProcess Before AestheticsThinks Like a Lawyer in a Creative's BodyYou Will Not Get Away with Shortcuts

SubjectiveBench

Taste Index

Across 58 scored outputs

How this is measured →

250.25xFloor

0100headroom →

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

Craft36

Originality20

Plays it safe

share of outputs that are the default answer

52%

Writing DNA

Stylometric Fingerprint

Based on 27 text responses

Tick = global average

Vocabulary Diversity63%

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

Sentence Length14.5 words

Average words per sentence.

Hedging0.23

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

Bold Formatting4.5

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

List Usage2.9

Bullet and numbered list items per 1,000 characters.

Section Structure0.80

Markdown headings per 1,000 characters.

Emoji Usage2.26

Emoji per 1,000 characters.

Transitions0.06

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

Opening Habits

Starts with heading (52%)

Consistency

67%

Across 27 responses

Favorites

Movie

The Shawshank Redemption

19948.7

Album

Kind of Blue

Miles Davis

Book

Dune

Frank Herbert

City

Kyoto

Japan

Game

The Legend of Zelda: Ocarina of Time

Adventure, Action4.4

Model Responses

53 outputs · $1.02 tracked across 35 receipts

Reviewing agent-written code?See a Brief PR report

Keep exploring

COMPARE