The Llama 4 Herd - Open Source Won?

5 Apr 2025 · 17 min
17:26

Meta released multiple Llama 4 models with open weights, including Scout and Maverick, with context windows reaching millions of tokens and Mixture-of-Experts architecture. The speaker compares performance to Gemini and GPT-4o while noting practical limits: massive VRAM requirements, licensing restrictions for large companies, and unproven video and coding capabilities.

Questions this video answers

April 5, 2025

Llama 4 Scout (aka Llama Force Scout), Llama 4 Maverick, Llama 4 Giant (Riesening) and Llama 4 Behemoth

17 billion active parameters; 16 experts (transcript also mentions total ~110 billion parameters with 17B active)

Claims of 10 million tokens (and later in the monologue a figure of 100 million tokens is mentioned)

Transcript reports 17 billion active parameters with 128 experts and later describes Maverick as ~400 billion total parameters with 17B active and 128 experts; context window claimed ~1 million tokens

Chapters

More from LARRUZ

8:49
youtube.com

OpenAI’s New AI Is Doing Things We Barely Understand

2d ago
5:38
youtube.com

DeepSeek Solved AI’s Memory Problem

4d ago
33:17
youtube.com

DeepMind's New AI Just Cracked The Code Of Life

1w ago
3:43
youtube.com

The Billion Dollar AI Advantage Is Disappearing

1w ago
5:14
youtube.com

Claude Opus 5.5 AI: An Incredible Leap Forward

3w ago
5:16
youtube.com

Yes, Jev Is Insane, But There's A Catch

3w ago
5:11
youtube.com

DeepSeek’s Insane New Architecture

3w ago
4:16
youtube.com

Claude Is Now Leaving Invisible Fingerprints In Its Text

1mo ago
9:18
youtube.com

I Never Thought I’d See This Happen

1mo ago
5:22
youtube.com

GPT-6 Astra Changes Everything

1mo ago
4:55
youtube.com

Claude Fable AI Is Much Stranger Than The Headlines Suggest

1mo ago
See all content