How to run agentic 35B models with only 8gb of vram (nvidia 4060ti)

"You need a 24 GB GPU for serious local LLMs in 2026." Everyone repeats this. It's not true anymore. In this video I go over new methods to run a 35B-parameter model on an RTX 4060 Ti 8 GB: • 41 tok/s at 16k context • 24 tok/s at 200k context. This approach uses a relatively common ryzen system with 64gb of vram and llama cpp. Let me know what you think in the comments below! X post - https://x.com/edgaraveloso/status/2050145746272321717 Rent an nVidia GPU on Vast AI (15% OFF) https://cloud.vast.ai/?ref_id=74601 For sponsorships or collaboration inquiries please contact aifluxcontact@proton.me Games footage: https://x.com/ZoldenGames OR https://t.co/u24MwZRsd7 Chapters 🎬: 0:00 - intro 0:32 - qwen 3.6 in smaller spaces 1:24 - post breakdown 2:00 - initial benchmarks 2:20 - how does it work? 2:30 - MoE offload 3:15 - suggested GPUs 3:35 - best context for agents 3:55 - flash attention 4:40 - 3070 also good? 5:44 - reddit users try this 6:28 - gpu specs 7:00 - my conclusion