AIbase
EN
Home
AI Models Finder
Llama 4 Maverick
L

Llama 4 Maverick

The Maverick large model is a native multimodal model capable of processing text and images. It adopts a Mixture of Experts (MoE) architecture with 17 billion active parameters and 128 experts, supporting a wide range of multimodal tasks such as conversational interaction, image analysis, and code generation. The model has a context window of 1 million tokens.
Intelligence(Medium)
Speed(Relatively Fast)
Input Supported Modalities
Yes
Is Reasoning Model
1,000,000
Context Window
1,000,000
Maximum Output Tokens
2024-08-01
Knowledge Cutoff
Go Compare

Pricing

￥1.22 /M tokens
Input
￥4.32 /M tokens
Output
￥2.78 /M tokens
Blended Price

Quick Simple Comparison

Input

Output

Llama 4 Scout
￥0.08
Llama 4 Maverick
￥0.17
Llama 3.2 Instruct 1B

Basic Parameters

Llama 4 MaverickTechnical Parameters
Parameter Count
400,000.0M
Context Length
1.0M tokens
Training Data Cutoff
2024-08-01
Open Source Category
Open Weights (Permissive License)
Multimodal Support
Text, Image
Throughput
638
Release Date
2025-04-05
Response Speed
164.43,645 tokens/s

Benchmark Scores

Below is the performance of Llama 4 Maverick in various standard benchmark tests. These tests evaluate the model's capabilities in different tasks and domains.
Intelligence Index
50.53
Large Language Model Intelligence Level
Coding Index
36.39
Indicator of AI model performance on coding tasks
Math Index
63.93
Capability indicator in solving mathematical problems, mathematical reasoning, or performing math-related tasks
MMLU Pro
80.9
Massive Multitask Multimodal Understanding - Testing understanding of text, images, audio, and video
GPQA
67.1
Graduate Physics Questions Assessment - Testing advanced physics knowledge with diamond science-level questions
HLE
4.8
The model's comprehensive average score on the Hugging Face Open LLM Leaderboard
LiveCodeBench
39.7
Specific evaluation focused on assessing large language models' ability in real-world code writing and solving programming competition problems
SciCode
33.1
The model's capability in code generation for scientific computing or specific scientific domains
HumanEval
87.9
Score achieved by the AI model on the specific HumanEval benchmark test set
Math 500 Score
88.9
Score on the first 500 larger, more well-known mathematical benchmark tests
AIME Score
39
An indicator measuring an AI model's ability to solve high-difficulty mathematical competition problems (specifically AIME level)
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase