Blog

Mixture of Experts (MoE) architectures are reshaping the landscape of large language models, offering efficiency gains that dense models can’t match. But these benefits come…

In AI, trust is fragile. One confidently wrong decision can derail adoption faster than a dozen wins can repair. Or as Tanmai Gopal puts it, “For serious work, one…

AI is reaching new dizzying heights with bigger models, more compute and a flood of pilots chasing transformation. Yet, inflated by compute hoarding and headline-driven…

We’re back with your monthly recap of what’s new across the PhariaAI stack. September’s updates bring sharper control, smoother customization and smarter workflows. All designed…

Currently, DeepSeek v3 is the most popular open-source large language model. The DeepSeek team recently introduced substantial inference-time optimizations, making the model surprisingly efficient to…

Transparency and control have been at the forefront of our innovation since we started to build LLM-based systems for the German and broader international market…

We’re excited to share something fresh from our research lab. For the bigger picture on why we’re building tokenizer-free models in the first place, check…

Your monthly drop of fresh features and all things Pharia. Here’s what to expect. From inference to infrastructure, and everything in between, these are the…

We signed the Code of Practice. We also helped shape it. Invited by the EU’s AI Office to contribute during its drafting, we supported a…