
Grok 3 Reasoning Beta
Grok 3, launched by xAI on February 17, 2025, is an advanced artificial intelligence model. Compared with Grok 2, its functions are significantly enhanced, and its performance has improved by an order of magnitude. Grok 3 is trained on a large dataset, including legal documents, etc., and utilizes the massive computing infrastructure of approximately 200,000 GPUs in the Memphis data center. The computing power used is ten times that of its predecessor. It has specialized models, such as Grok 3 Reasoning and Grok 3 Mini Reasoning, to solve complex problems and performs excellently in benchmark tests such as the AIME in mathematics and the GPQA in doctoral-level science.
Intelligence(Medium)
Speed(Slow)
Input Supported Modalities
Yes
Is Reasoning Model
1,000,000
Context Window
8,000
Maximum Output Tokens
2024-11-17
Knowledge Cutoff
Pricing
¥21.6 /M tokens
Input
¥108 /M tokens
Output
- /M tokens
Blended Price
Quick Simple Comparison
Grok 3 Reasoning Beta
¥3
Grok 3
¥3
Grok 3 mini Reasoning (Low)
¥0.3
Basic Parameters
Grok 3 Reasoning BetaTechnical Parameters
Parameter Count
Not Announced
Context Length
1.0M tokens
Training Data Cutoff
2024-11-17
Open Source Category
Proprietary
Multimodal Support
Text, Image
Throughput
100
Release Date
2025-02-17
Response Speed
0 tokens/s
Benchmark Scores
Below is the performance of Grok 3 Reasoning Beta in various standard benchmark tests. These tests evaluate the model's capabilities in different tasks and domains.
Intelligence Index
56.1
Large Language Model Intelligence Level
Coding Index
-
Indicator of AI model performance on coding tasks
Math Index
-
Capability indicator in solving mathematical problems, mathematical reasoning, or performing math-related tasks
MMLU Pro
-
Massive Multitask Multimodal Understanding - Testing understanding of text, images, audio, and video
GPQA
84.6
Graduate Physics Questions Assessment - Testing advanced physics knowledge with diamond science-level questions
HLE
-
The model's comprehensive average score on the Hugging Face Open LLM Leaderboard
LiveCodeBench
79.4
Specific evaluation focused on assessing large language models' ability in real-world code writing and solving programming competition problems
SciCode
-
The model's capability in code generation for scientific computing or specific scientific domains
HumanEval
-
Score achieved by the AI model on the specific HumanEval benchmark test set
Math 500 Score
-
Score on the first 500 larger, more well-known mathematical benchmark tests
AIME Score
93.3
An indicator measuring an AI model's ability to solve high-difficulty mathematical competition problems (specifically AIME level)
GPT 5 Mini
openai

¥1.8
Input tokens/million
¥14.4
Output tokens/million
400k
Context Length
GPT 5 Standard
openai

¥63
Input tokens/million
¥504
Output tokens/million
400k
Context Length
GPT 5 Nano
openai

¥0.36
Input tokens/million
¥2.88
Output tokens/million
400k
Context Length
GPT 5
openai

¥9
Input tokens/million
¥72
Output tokens/million
400k
Context Length
GLM 4.5
chatglm

¥0.43
Input tokens/million
¥1.01
Output tokens/million
131k
Context Length
Gemini 2.0 Flash Lite (Preview)
google

¥0.58
Input tokens/million
¥2.16
Output tokens/million
1M
Context Length
Gemini 1.0 Pro
google

¥3.6
Input tokens/million
¥10.8
Output tokens/million
33k
Context Length
Qwen2.5 Coder Instruct 32B
alibaba

¥0.65
Input tokens/million
¥0.65
Output tokens/million
131k
Context Length