Meta Launches Llama-4 AI Suite: Scout, Maverick & Behemoth to Rival ChatGPT, Gemini

Meta, led by CEO Mark Zuckerberg, has intensified its efforts in the competitive artificial intelligence landscape by launching its next-generation Llama-4 (Large Language Model) suite. Announced on April 6, 2025, the suite includes three powerful AI models – Scout, Maverick, and Behemoth – each tailored to different multimodal and reasoning capabilities. This strategic release positions Meta to directly compete with leading AI platforms like OpenAI’s ChatGPT and Google’s Gemini, while also responding to emerging global players like China’s DeepSeek.

Key Highlights and Features

• Launch Date: April 6, 2025

Models Introduced

• Scout – Meta claims it is the best multimodal model available.
• Maverick – Described as Meta’s flagship AI assistant.
• Behemoth – The most powerful LLM developed by Meta, intended to act as a teacher model.

Capabilities

• Handle advanced reasoning across text, images, and video.
• Perform document summarization, code generation, and visual processing.
• Support for multimodal tasks with high precision and personalization.

Performance Claims

• Outperforms Gemma 3, Gemini 2.0, Flash Lite, Mistral 3.1.
• Maverick reportedly exceeds GPT-4o and Gemini 2.0 in coding, reasoning, and image-based tasks.
• Scout hailed as the most accurate and capable multimodal AI model.

Availability

• Scout and Maverick are freely available on Meta’s website.
• Behemoth is in preview stage, not yet publicly released.

Meta’s AI Infrastructure & Investment Plans

• Investment Target for 2025: $65 billion for AI development and deployment.

Key Projects

• A $10 billion data center in Louisiana is under construction.
• Acquisition of new high-performance chips to power AI operations.
• Hiring more talent across AI research, development, and deployment teams.

Goal

• To make Meta AI the most widely used AI assistant globally by end of 2025.
• Seamlessly integrate Meta AI into social media platforms.
• Expand to AI-powered smart glasses and wearable technology.

Response to Global AI Competition

• Pressure from DeepSeek
• China’s DeepSeek models (R1 and V3) surpassed Llama-2 in benchmarks.
• Meta responded by accelerating Llama-4’s development with internal “AI war rooms”.
• Behemoth designed to reverse-engineer efficiency and cost innovations seen in DeepSeek.

Summary/Static	Details
Why in the news?	Meta Launches Llama-4 AI Suite: Scout, Maverick & Behemoth to Rival ChatGPT, Gemini
New AI Models	Scout, Maverick, Behemoth
Model Type	Llama-4 (Large Language Model family)
Key Feature	Scout Claimed as world’s best multimodal AI
Key Feature	Maverick Flagship AI assistant; outperforms GPT-4o and Gemini 2.0
Key Feature	Behemoth Meta’s most powerful LLM; acts as a “teacher” model
Applications	Text, image, video reasoning; summarization; code generation
Performance Benchmarks	Beats Gemini 2.0, Gemma 3, Flash Lite, Mistral 3.1
Availability	Scout & Maverick – Free on Meta’s website; Behemoth – In preview stage
Investment for AI (2025)	~$65 billion
Infrastructure Projects	$10 billion data center (Louisiana), high-performance chip acquisitions
AI Integration Plan	Embed Meta AI across social platforms; develop smart glasses
Competition	Targeted OpenAI (ChatGPT), Google (Gemini), DeepSeek (China)