SupraLabs_

Small weights.
Huge potential.

We're democratizing AI by training high-performance models on consumer hardware. 100% Open Source. 0% Slop.

// Latest_Updates

2026-05-16

Supra Mini v5 8M is out!

This is scaling up so much 👀

#scalinglaws #slm #open-source

2026-05-15

Supra Mini v4 2M is out!

Take a look! It looks coherent!

#smallmodel #llama #open-source

2026-05-14

Supra Mini v3 0.5M is out!

And it's huge! Read all about it here...

#new #model #open-source

2026-05-12

1Bit quantization: a sensitive area for tiny models

1Bit quantization, a 50/50 chance area for tiny models...

#research #1bit #llama #open-source #1.58b

2026-05-12

Why we chose Llama over Mistral, DeepSeek, Qwen and Mamba for Supra Mini v3

Researching the best architecture for small scale language models...

#research #architecture #llama #open-source #tinyml

2026-05-12

Our new model is out: Supra Mini v2 0.1M

Check it out! It's cool!

#new #model #open-source

2026-05-12

Our first model ever is there: Supra Mini 0.1M

This is our first model at SupraLabs. Free. Tiny. Open Source.

#model #new #open-source

2026-05-11

We founded SupraLabs: Real Open Source AI

Today marks the beginning of our journey. No corporate gatekeeping, just pure optimization on accessible hardware.

#announcement #roadmap #open-source