Probably one of cheapest and performant local setups you can get ( ~ 1600€ )
retained text
i7-14700KF , 64 GB RAM Radeon RX 9070XT 16GB [https://github.com/niko1221/strata](https://github.com/niko1221/strata) for inference Qwen3.8-Flash-Next-GSQ-RCO-Q2\_0 model Runs at 90-70 tokens/sec :)
document fetched 2026-10-24 10:42 · canonical https://reddit.com/r/LocalLLaMA/comments/1xdr8pe/probably_one_of_cheapest_and_performant_local