Meta to open source next-gen AI Models: What to expect from upcoming releases

Meta, the social media giant, is gearing up to launch new AI models developed by its Superintelligence Labs (MSL) team, which is led by Chief AI Officer Alexandr Wang. According to the Axios report, Meta plans to open-source versions of these AI models, allowing developers and companies to access, use, and build on them.

The report further highlights that before releasing its new models publicly, Meta may plan to keep some parts private to make sure the models don’t create new safety risks. Previously, there have been reports surrounding Meta’s plan to stop or reduce this open approach, but it may adopt a more balanced strategy by bringing tighter control over sensitive components.

The upcoming AI models by Meta are an effort to regain competitive ground after Llama 4 underperformed, and as OpenAI and Anthropic signal upcoming releases.

 In addition, the company is more focused on building modes that are widely adopted by consumers. Therefore, it aims to distribute its AI models and tools more broadly. On the other hand, rivals like OpenAI and Anthropic are more focused on building models to align with governments and enterprise requirements.

To back its AI strategy, Meta has recently made a massive hardware investment and is spending heavily on infrastructure, especially by buying the NVIDIA H100 chips. These chips provide the huge computing power required to train advanced AI models. With its hardware power, Meta can train bigger and more powerful models.

While Meta is racing towards consumer-centric products and AI models, rivals are more focused on enterprise-based AI tools. Recently, Anthropic’s Claude Cowork made headlines for its tools and capabilities that could work across divisions and sectors. Whereas OpenAI is also making significant upgrades to its Codex platform, enhancing its ability to assist developers with more advanced coding, automation, and enterprise-level workflows.