SupraLabs launches open-source small AI models, starting with 2M-parameter mini
Small but powerful: SupraLabs' 2M-parameter model runs on edge devices—free on Hugging Face.
SupraLabs, a new open-source AI initiative founded by a Reddit community, has launched its first batch of small language models on Hugging Face. The flagship model, Supra-Mini-v4-2M, packs just 2 million parameters—tiny by today's standards—but aims to deliver strong performance for edge devices. The team emphasizes training, fine-tuning, and exploring small models to make AI accessible to everyone, not just those with massive cloud budgets.
Upcoming models include StorySupra 10M, a storytelling SLM optimized for edge devices, and Supra Mini v5 5M, boasting cutting-edge performance despite its small size. SupraLabs offers a blog for updates and invites community contributions. This project is a direct challenge to the trend of ever-larger models, proving that small, open-source AI can be both capable and widely deployable on consumer hardware.
- SupraLabs released Supra-Mini-v4-2M, a 2M-parameter open-source model on Hugging Face.
- Upcoming models include StorySupra 10M (storytelling) and Supra Mini v5 5M with 'great results.'
- All models are free, open-source, and optimized for edge devices and local deployment.
Why It Matters
Small, open-source models let professionals run AI locally on low-cost hardware, democratizing access to powerful language AI.