GR

Grok 4

Reasoning modelGrok 4

Grok 4

Release date: 2025-07-10Updated: 2025-08-09 22:51:233,149
Live demoGitHubHugging FaceCompare
Parameters
Not disclosed
Context length
256K
Chinese support
Supported
Reasoning ability

Grok 4 is an AI model published by xAI, released on 2025-07-10, for Reasoning model, and 256K context length, under the 不开源 license, with a 98.80 score on AIME2025.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Grok 4

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
256K tokens
Max output length
256K tokens
Model type
Reasoning model
Modality (in / out)
Text, Image → Text
Release date
2025-07-10
Model file size
No data
MoE architecture
No
Total params / Active params
No data / N/A
Knowledge cutoff
No data
Grok 4

Open source & experience

Code license
不开源
Weights license
不开源
GitHub repo
GitHub link unavailable
Hugging Face
Hugging Face link unavailable
Grok 4

Official resources

Paper
DataLearnerAI blog
Grok 4

API details

API speed
3/5
No public API pricing yet.
Grok 4

Benchmark Results

Grok 4 currently shows benchmark results led by IMO 2024 (1 / 10, score 23.20), IMO 2025 (1 / 9, score 29.20), MMLU Pro (14 / 126, score 87). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking

General Knowledge

8 evaluations
Benchmark / mode
Score
Rank/total
87
14 / 126
87
39 / 179
66.70
29 / 65
LiveBench
Standard Mode
62.02
59 / 115
38.60
55 / 159
38.60
55 / 159
25.40
88 / 159
15.90
34 / 59

Coding and Software Engineer

2 evaluations
Benchmark / mode
Score
Rank/total
82
25 / 120
58.60
79 / 108

Math and Reasoning

9 evaluations
Benchmark / mode
Score
Rank/total
98.80
13 / 106
91.70
36 / 106
46.70
4 / 16
23.30
10 / 16
29.20
1 / 9
23.20
1 / 10
12.10
22 / 60
2.10
56 / 80

AI Agent - Tool Usage

1 evaluations
Benchmark / mode
Score
Rank/total

常识推理

1 evaluations
Benchmark / mode
Score
Rank/total
Simple Bench
Thinking Mode
60.50
15 / 63

Agent Level Benchmark

2 evaluations
Benchmark / mode
Score
Rank/total
79.60
7 / 59

Compare with other models

No curated comparisons for this model yet.

Want a custom combination? Open the compare tool

Grok 4

Publisher

Grok 4

Model Overview

Grok 4 is an AI model published by xAI, released on 2025-07-10, for Reasoning model, and 256K context length, under the 不开源 license, with a 98.80 score on AIME2025.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code