Grok 4Overview
A detailed comparison of Grok 4, focusing on their key differences and practical applications.
Performance Benchmarks
Comparative analysis across standard metrics
16 Benchmarks
| Aider-Polyglot | 79.6% |
| AIME 2025 | 98.8% |
| ARC-AGI-1 | 66.7% |
| EQ-Bench 3 | 119320.0% |
| ARC-AGI-2 | 16.2% |
| Confabulations | 1241.0% |
| Creative Story Writing | 769.0% |
| Elimation Game | 599.6% |
| GPQA Diamond | 87.5% |
| HMMT 2025 | 93.9% |
| Humanity's Last Exam | 38.6% |
| NYT Connections | 92.4% |
| Thematic Generalisation | 188.0% |
| USAMO 2025 | 37.5% |
| MathArena Apex | 2.1% |
| Vending Bench 2 | 199946.0% |
Grok 479.6%
Grok 498.8%
Grok 466.7%
Grok 4119320.0%
Grok 416.2%
Grok 41241.0%
Grok 4769.0%
Grok 4599.6%
Grok 487.5%
Grok 493.9%
Grok 438.6%
Grok 492.4%
Grok 4188.0%
Grok 437.5%
Grok 42.1%
Grok 4199946.0%
Pricing Analysis
Price comparison per million tokens
Input $/M-
Output $/M-
Total (1:3) $/M-
API Availability & Providers
How many providers expose each model based on current pricing configuration.
Subscription Plans
Plans that include each selected model, grouped by organisation.
Grok 4
2 plans
SuperGrokxAI
$30.00 / monthly · $300.00 / yearly
$300.00 / monthly · $3000.00 / yearly
Context Window
Maximum input and output token capacity
All Models
Input Context Length256K
Output Context Length256K
License Type
Model usage and distribution terms
Proprietary
Release Timeline
Model release chronology
Grok 4 was released on 10 Jul 2025.
Knowledge Cutoff Timeline
Not enough models with knowledge cutoff info
Model Comparison
| General Information | |
| Context Window | Input: 256,000 Output: 256,000 |
| Multimodal | |
| Parameters | - |
| Training Tokens | - |
| License | Proprietary |
| Knowledge Cutoff | Nov 2024 |
| Operational Metrics | |
| Cost per 1M Tokens | Input: - Output: - |
| Latency | - |
| Throughput | - |
| Benchmarks | |
| AIME 2025 | |
| ARC-AGI-1 | |
| ARC-AGI-2 | |
| Aider-Polyglot | |
| Confabulations | |
| Creative Story Writing | |
| EQ-Bench 3 | |
| Elimation Game | |
| GPQA Diamond | |
| HMMT 2025 | |
| Humanity's Last Exam | |
| MathArena Apex | |
| NYT Connections | |
| Thematic Generalisation | |
| USAMO 2025 | |
| Vending Bench 2 | |
General Information
Context Window:
Input:256,000
Output:256,000
Multimodal:
Parameters:-
Train Tokens:-
License:Proprietary
Knowledge Cutoff:Nov 2024
Operational Metrics
Cost per 1M Tokens:
Input:-
Output:-
Latency:-
Throughput:-
Benchmarks
AIME 20250.99
ARC-AGI-10.67
ARC-AGI-20.16
Confabulations12.41
EQ-Bench 31,193.2
GPQA Diamond0.88
HMMT 20250.94
MathArena Apex0.02
NYT Connections0.92
USAMO 20250.38
Vending Bench 21,999.46