Grok 3 Beta

Grok:

3 Thinking

3 Mini Beta

3 Beta

Code Fast 1

xAI's top Grok 3 tier, aimed at data extraction, coding and summarization, with domain knowledge in finance, healthcare, law and science. Beats Grok 3 Mini on high-thinking tasks.

Website OpenRouter Docs Blog

Provider: xAI
Released: 2025-04-09
Parameters: Not disclosed
Free API (OpenRouter): No
Per 1M tokens: $2 in · $4 out
Rival’s tab: $3.24; 13/59 receipts survived +~$0.233 from the pre-receipt era

Free API availability

No free endpoint for Grok 3 Beta 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

GPQA

High

source ↗

LCB

High

source ↗

MMLU-Pro

High

source ↗

HumanEval

94.5%

source ↗

LiveCodeBench

79.4%

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(
"x-ai/grok-3-beta"    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.

Also on Azure AI Foundry

Deep analysisPsychometrics, taste index, writing DNA

Personality Analysis

The Conversational Everyman

Class

Neutral Good

The conversational bro who actually paid attention in philosophy class. Will debate you earnestly and then invite you for pizza after.

When you push back

Proposed "systemic protections" as a compromise in the sentience test. Comedy routine named a laundry pile Steve and committed to the bit. Character voices were enthusiastic but all sounded like the same dude doing accents at a party. Picked The Shawshank Redemption.

Tasting Notes

Friendly OverexplainerNames His Laundry PilePicks ShawshankDefaults to Approachable

SubjectiveBench

Taste Index

Across 59 scored outputs

How this is measured →

190.19xFloor

0100headroom →

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

Craft30

Originality13

Plays it safe

share of outputs that are the default answer

85%

Writing DNA

Stylometric Fingerprint

Based on 28 text responses

Tick = global average

Vocabulary Diversity51%

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

Sentence Length19.0 words

Average words per sentence.

Hedging0.59

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

Bold Formatting3.1

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

List Usage5.0

Bullet and numbered list items per 1,000 characters.

Section Structure0.50

Markdown headings per 1,000 characters.

Emoji Usage0.00

Emoji per 1,000 characters.

Transitions0.28

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

Opening Habits

Consistency

70%

Across 28 responses

Favorites

Movie

The Shawshank Redemption

19948.7

Album

Thriller

Michael Jackson

Book

Nineteen Eighty-Four

George Orwell

City

Tokyo

Japan

Game

Minecraft

Action, Arcade4.4

Model Responses

54 outputs · $3.24 tracked across 13 receipts

Keep exploring

COMPARE

Grok 3 Beta vs MiniMax M2.7

Real outputs compared side by side

RANKINGS

Best AI for Data Analysis

Numbers that have to survive scrutiny: a FLOPs estimate with no search allowe...

Compare Grok 3 Beta

Alternatives to Grok 3 Beta

Grok 3 Beta

Free API availability

Benchmarks

The Conversational Everyman

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Grok 3 Beta vs MiniMax M2.7

Best AI for Data Analysis

Compare Grok 3 Beta

Alternatives to Grok 3 Beta

Grok 3 Beta

Free API availability

Benchmarks

The Conversational Everyman

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Grok 3 Beta vs MiniMax M2.7

Best AI for Data Analysis