https://huggingface.co/aquif-ai/aquif-3.5-Nano-1B
#1574
by
aquiffoo
- opened
New model from aquif AI! built on Qwen3-1.7B, but non-thinking
It's queued! :D
Wow looks amazing for a 1.7B model based on your posted benchmarks. I never realized how even tiny models are now getting so good at MMLU. I still remember from a few years ago when such MMLU cores were only achieved by the largest state of the art models. Just as reference original Llama 65B was 63.4% at MMLU at 5-shot.
You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#aquif-3.5-Nano-1B-GGUF for quants to appear.