DEV Community

Cover image for Benchmarking 2-bit GGUFs fairly: paired windows, and the NaN trap on RTX 50 GPUs
Genrius AI
Genrius AI

Posted on Originally published at genrius-ai.github.io

Benchmarking 2-bit GGUFs fairly: paired windows, and the NaN trap on RTX 50 GPUs

Try it

Live chat with Qwen3.8-27B and Qwen3.5-2B v2 in your browser: https://huggingface.co/spaces/harshit-genrius/pingala8-qwen35-live

Originally published at https://genrius-ai.github.io/pingala8-llm/benchmarking-2-bit-ggufs-paired-windows.html

This article was written with the help of AI; every number comes from our measured logs and the paper.

Top comments (0)