xAI logo

Grok-4 Heavy

Multimodal
xAI

Grok 4 Heavy is a multi-agent version of Grok 4, released alongside the standard model in summer 2025. The system runs multiple Grok 4 agents in parallel that work independently on tasks and then combine solutions, similar to a study group. Agents share discovered insights and techniques, and the system intelligently combines their work instead of simple majority voting. Uses approximately 10x more compute at test time than regular Grok 4.

Key Specifications

Parameters
-
Context
-
Release Date
July 8, 2025
Average Score
83.0%

Timeline

Key dates in the model's history
Announcement / Last Update
July 8, 2025
Today
March 25, 2026

Technical Specifications

Parameters
-
Training Tokens
-
Knowledge Cutoff
December 1, 2024
Family
-
Capabilities
MultimodalZeroEval

Benchmark Results

Model performance metrics across various tests and benchmarks

Reasoning

Logical reasoning and analysis
GPQA
AccuracySelf-reported
88.0%

Other Tests

Specialized benchmarks
AIME 2025
AccuracySelf-reported
100.0%
HMMT25
AccuracySelf-reported
97.0%
LiveCodeBench
AccuracySelf-reported
79.0%
HLE
AccuracySelf-reported
51.0%

License & Metadata

License
proprietary
Announcement Date
July 8, 2025
Last Updated
July 8, 2025

Similar Models

All Models

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.