Contributors › magnusfroste's local-LLM benchmark configs
magnusfroste's local-LLM benchmark configs
Every tracked benchmark config measured by magnusfroste: model, quant, backend, tokens per second, source-linked and trust-tiered.
Snapshot 2026-10-15 · 394 configs on AI boxes and 423 on GPU cards · 123 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.
4 configs · 4 models · 49 tok/s best single-stream · magnusfroste on github.
Configs measured by magnusfroste
| Model | Hardware | Quant | Backend | Decode | Context | Trust | Source |
|---|---|---|---|---|---|---|---|
| Qwen3.6-35B-A3B | DGX Spark | FP8 | vllm | 49 tok/s | 256000 | Extracted, source-linked | source |
| Qwen3.8-27B | DGX Spark | NVFP4 | vllm | 11 tok/s | 256000 | Extracted, source-linked | source |
More: all configs · hardware · news · contributors · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.