Live
Uluka Dango
Compressed from Google's Gemma 4 26B. Formerly called Uluka 26B.
51.6 GB to 8.7 GB, 5.9× smaller
- Averages 69.9 on four benchmarks against 70.4 for Google's 4-bit version, using 44% less memory
- Reads text and images
- Runs on a single NVIDIA GPU with 12 GB or more
- Free to use in Uluka chat, and the model behind Uluka Code and the Uluka app
See the results · Try it in Uluka chat
Paused
Uluka Tiramisu
Our larger model, built on open-weight models
- For harder questions, longer writing and code
- Part of Uluka Plus, alongside unlimited Uluka Dango and Uluka Code
- Paused for now while we prepare its launch
See Uluka Plus
Proof of concept
Uluka 7B
Compressed from AI2's OLMoE-1B-7B
13.8 GB to 2.8 GB, 4.9× smaller
- Keeps about 98% of the original's accuracy across seven benchmarks, corrected for chance
- Runs on a laptop with no dedicated GPU: about 6 tokens a second on an Intel Core Ultra 5
- We built it to check the method also works on a second model design. Speed was not a goal.
Planned
Uluka 266B
To be compressed from Thinking Machines' Inkling-Small
Target: about 80 GB on a single 96 GB GPU
- Inkling-Small is about 10 times the size of Gemma 4 26B, with an Apache-2.0 license. It reads text, images and audio.
- Its official builds need two to four 96 GB GPUs. We aim to fit it on one. We project 59% to 80% potential savings on GPU cost.
- Not built yet. These are targets, not results.
How larger models save more