I’d like to request evaluation of a new Hugging Face model:
👉 https://huggingface.co/rfcoder0/qwen3-4b-custom-Sile
• Model size: 4B parameters
• Training hardware: RTX 3060 + RTX 3070 (consumer GPUs)
• Performance: Rivals Qwen-3 8B
• Intended use: General-purpose chat / instruct
This is a custom 4B model designed to show that high performance is possible with modest hardware. Would be great to see it included in your evaluation sweeps/leaderboards for comparison.
Thank you!