Running MiniMax H3 Locally: Extremely High VRAM and RAM Usage Reported
Full_Astronomer_5438 · reddit · 2026-08-03
Reddit users report extremely high resource consumption when running the MiniMax H3 model locally in ComfyUI. The model causes severe inference spikes and drives GPU temperatures to unusually high levels.
Even with a pruned INT8 quantization and NVFP4 text encoder, the model quickly consumes 10-30GB of system page file (virtual memory) depending on the video length (5-15 seconds).
Related event: MiniMax H3 Local Tests: Benchmark Results and Comparisons Across GPUs(88 posts)→
More from Infra
- Trading 2 RTX 5090s for 4 R9700s: does doubled VRAM pay off? — poofph · 2026-10-16
- Running PolyStrata + Cline coding agent on a $0-fancy laptop at ~90 tok/s — rpwoerk · 2026-10-16
- TensorFold 1.0.12 runs 120B vision model on 1-2 DGX Sparks, adds Mac image input — HankYeomans · 2026-10-16
- vLLM brings frontier model local deployment to NVIDIA RTX Spark PCs — vllm_project · 2026-10-16
- NVIDIA signals $2B in new European AI deals at GTC Berlin, totaling ~$10B in a year — CurieuxExplorer · 2026-10-16
- Making Your Data Ready for Agentic AI: trust first, then semantics, then writes — bibryam · 2026-10-16