Xai logo

Grok 3 mini Reasoning (high) – Benchmarks, Pricing & Intelligence Analysis

Not available on LLMBase

Grok 3 mini Reasoning (high) is not currently available in LLMBase Chat or the LLMBase Inference API. Benchmarks, pricing, and model details remain available for research and comparison.

Input Price$0.30/1M tokens
Output Price$0.50/1M tokens
Intelligence22.5
CodingN/A

Specifications

Technical details and pricing.

ProviderXai
Context WindowN/A
Release DateFeb 19, 2025
ModalitiesText
AvailabilityNot available

Benchmarks

12 benchmark scores from Artificial Analysis.

GPQA79.1%
MMLU Pro82.8%
HLE11.1%
LiveCodeBench69.6%
MATH 50099.2%
AIME 202584.7%
AIME93.3%
SciCode40.6%
LCR50.3%
IFBench45.9%
Tau290.4%
TerminalBench Hard17.4%

Composite Indices

Higher is better; speed and price are normalized

Standard Benchmarks

Only benchmarks with data are shown

EU-Hosted Inference API

Power your AI projects with the best open-source models.

Drop-in OpenAI-compatible API. No data leaves Europe.

Explore Inference API

MiniMax

MiniMax M3

$0.40 / $1.40

per M tokens

Z.ai

GLM 5.2

$1.40 / $4.40

per M tokens

MoonshotAI

Kimi K2.7 Code

$0.95 / $4.00

per M tokens

DeepSeek

DeepSeek V4 Pro

$1.80 / $3.60

per M tokens

Frequently Asked Questions

What is Grok 3 mini Reasoning (high) good for?

Grok 3 mini Reasoning (high) currently has benchmark and pricing data on this page. Runtime metadata will appear later when the model catalog entry becomes available.

How much does Grok 3 mini Reasoning (high) cost?

Pricing is based on usage. Current rates are $0.30/1M tokens for input and $0.50/1M tokens for output.

Is Grok 3 mini Reasoning (high) available in LLMBase Chat yet?

Not yet on this page. We will add chat and inference actions automatically once an exact catalog entry exists for Grok 3 mini Reasoning (high).

Does Grok 3 mini Reasoning (high) support images or audio?

Grok 3 mini Reasoning (high) focuses on text-based tasks.

Benchmarks and pricing use Artificial Analysis where available. Catalog specs are used as a fallback.