Rent a GPU for ~€20/month: Gemma 4 31B hits 41 TPS on GeForce NOW via LM Studio
evilsocket · x · 2026-10-11
A dev demonstrated running LM Studio on GeForce NOW's Ultimate tier, getting Gemma 4 31B at 41 tokens/s. evilsocket points out the Ultimate subscription costs only about €20/month — effectively a cheap way to rent a high-end GPU for local LLM inference instead of buying hardware.
More from Infra
- Firmus Grid pulls ASX listing at $30B valuation as Altman says now ill-advised to IPO — shashib · 2026-10-11
- clef-flash distilled into 0.6B model on Apple Neural Engine triages 1,000 tickets in 57s — art_zucker · 2026-10-11
- B300 GPU Prices Up 38% Since Token Index Peak: Open Models Boost AI Infra — GavinSBaker · 2026-10-11
- Nvidia-backed data center operator pulls $5B IPO, raises privately instead — YvesMulkers · 2026-10-11
- Mixing an RTX 5080 with a 3080Ti to offload CLIP/VAE for local video gen models — MonkeyCartridge · 2026-10-11
- Dev's take: Mac beats Linux laptops for 24/7 local AI like Qwen 30B — TheZachMueller · 2026-10-11