Meta may charge for its next Avocado AI model

B
Baseinsider Team
Author
Articles
Nov 28, 2025
1 min read
157 views
Meta may charge for its next Avocado AI model
Meta is working on a new Avocado AI model that may not be freely open, signaling a shift away from the fully open source posture that framed its earlier Llama releases.

Meta may charge for its next Avocado AI model

The piece looks at reports that Meta is developing a new model code named Avocado and considering charging for access. That would be a notable change from the company earlier positioning around open source style releases for Llama, and suggests Meta wants more direct control over how its most capable models are used and monetized.

Main ideas in the article

  • Avocado could be a higher end model that sits behind paid licensing or usage fees.
  • Meta leadership is rethinking how far to push open releases after mixed results with Llama 4 and intense competition.
  • The company is investing heavily in AI talent and infrastructure, and a paid model could be one way to justify that spending.

Read the original article on The Verge.

Related Articles

Amazon appoints longtime AWS exec Peter DeSantis to lead new AI org

Amazon appoints longtime AWS exec Peter DeSantis to lead new AI org

Amazon has appointed longtime AWS leader Peter DeSantis to lead a new AI-focused division aimed at advancing its AI models, silicon technology, and quantum computing efforts. This move reflects Amazon’s growing commitment to AI innovation and infrastructure investments.

Dec 08, 2024 2 min read
MIT researchers use AI to uncover atomic defects in materials

MIT researchers use AI to uncover atomic defects in materials

MIT scientists have created an advanced AI model that can identify and quantify atomic defects in materials, a breakthrough for improving products like semiconductors, solar cells, and batteries. Leveraging data from noninvasive neutron-scattering, their system can pinpoint up to six types of defects at once, covering a broad range of materials. This approach could lead to greater control over material properties, though there are still challenges integrating it into existing manufacturing workflows.

Apr 01, 2026 2 min read
How the Internet Broke Everyone’s Bullshit Detectors

How the Internet Broke Everyone’s Bullshit Detectors

As AI-generated media and restricted satellite imagery spread online, traditional ways to verify digital content are faltering. The rise of fast-moving synthetic information—often indistinguishable from real events—means audiences must rely more on careful checks and skepticism. Until more robust tools emerge, the best defense is to pause and question before sharing.

Apr 11, 2026 3 min read