Grok 4.5
Parameters
Undisclosed
Context Length
500K
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
8 Jul 2026
Knowledge Cutoff
Feb 2026
API Pricing (per 1M)
Input: $2.00 · Output: $6.00
Evaluation Benchmarks
Rank
#22
| Benchmark | Score | Rank |
|---|---|---|
Graduate-Level QA | 0.93 | 8 |
Agentic Coding | 0.56 | 20 |
Reasoning | 0.87 | 22 |
Agent Arena | 0.03 | 23 |
Mathematics | 0.91 | 24 |
LiveBench Average | 0.76 | 25 |
General | 0.76 | 26 |
Web Development | 1555 | 27 |
Coding Index | high 0.72 | 31 |
Data Analysis | 0.73 | 31 |
Agentic Index | high 0.42 | 36 |
General Text | 1468 | 39 |
Intelligence Index | high 0.39 | 40 |
Coding | 0.69 | 46 |
Rankings
Overall Rank
#22
Coding Rank
#29
About Grok 4.5
Grok 4.5 is a frontier multimodal reasoning model from xAI offering strong performance across coding, STEM, and deep analytical tasks. It is equipped with a 500K context window and handles text, image, and file inputs.
Technical Specifications
Architecture specifications are undisclosed for proprietary models.
Attention
Attention Structure
Multi-Head Attention
Attention Heads
Undisclosed
Key-Value Heads
Undisclosed
Attention Head Dimension
Undisclosed
Position Embedding
Absolute Position Embedding
RoPE Theta
Undisclosed
Sliding Window Attention
Undisclosed
Sliding Window Size
Undisclosed
Sliding Window Ratio
Undisclosed
Linear Attention
Undisclosed
Linear Attention Ratio
Undisclosed
Normalization
Undisclosed
Activation Function
Undisclosed
Dimensions
Hidden Dimension Size
Undisclosed
Number of Layers
Undisclosed
FFN Intermediate Size (Dense)
Undisclosed
Multi-Token Prediction Heads
Undisclosed
Tokenizer
Vocabulary Size
Undisclosed
About Grok 4
xAI's frontier intelligence models trained with reinforcement learning at unprecedented scale using the 200,000 GPU Colossus cluster. Grok 4 series demonstrates state-of-the-art performance in reasoning, coding, and multimodal understanding with native tool use capabilities. Features real-time search integration across X and the web, advanced reasoning through scaled RL training, and industry-leading performance on academic benchmarks. Designed for both immediate responses and extended thinking modes with vision capabilities.