Contributors › scouzi1966's local-LLM benchmark configs

scouzi1966's local-LLM benchmark configs

Every tracked benchmark config measured by scouzi1966: model, quant, backend, tokens per second, source-linked and trust-tiered.

Snapshot 2026-09-30 · 254 configs · 116 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.

1 configs · 1 models · 76.4 tok/s best single-stream · scouzi1966 on huggingface.

Configs measured by scouzi1966

ModelQuantBackendDecodeContextTrustSource
Qwen3.8-Flash-NextMLX-4BITmlx (speculative decode, mtp)76.4 tok/sExtracted, source-linkedsource

More: all configs · hardware · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.