← Back to feed

I benchmarked 22 local LLMs looking for the Qwen 3.8 35B we didn't get

AI · Data · 2026-11-19 19:40

The author built SpiruxBench to benchmark 22 quantized local LLMs suitable for gaming PCs, specifically investigating whether a Qwen 3.6 fine-tune can serve as a viable substitute for the unreleased Qwen 3.8 MoE model.

Source: sevangelatos.github.io


Auto-generated summary for reference only. RSS