LlaMa 4 Maverick & Scout

Meta has released the most powerful OpenSource model of 2025, the LlaMa 4 Maverick and the less one, the LlaMa 4 Scout.

Meta has always been behind in the AI ​​race. It always took a long time to introduce new features, and when it did, it was already months late and with poor quality. It seems that NVIDIA taught Meta how to make a good AI model, as it did with Nemotron.

Now, LlaMa 4 Maverick is finally ready, with great performance and best of all: it is not behind in the race, presenting updated features and good quality.

Will they stay in the race or will DeepSeek soon overtake them?

https://ai.meta.com/blog/llama-4-multimodal-intelligence/