What GPU Do I Need to Run MiniMax H3 Locally?

Overview

There is no answer yet, because you cannot run MiniMax H3 locally at the moment. As of early August 2026 the model is API-only: it was announced on July 31, 2026 as open-weight, but the weights have not shipped and MiniMax has published no hardware or VRAM requirements. Source note: Verified 2026-08-06 against the MiniMax official H3 blog, MiniMax Video Generation API docs, and Hugging Face MiniMax-H3 model page.

So the honest current guidance is that you do not need any GPU. Using MiniMax H3 today means calling a hosted API, where generation runs on the provider’s hardware and you pay per use rather than buying a card.

When the open weights are released, expect a capable modern GPU with generous VRAM to be necessary — video models are memory-hungry, and a 2K, audio-generating omni-modal model is unlikely to be light — but treat any specific number as unverified until MiniMax publishes official requirements. Do not plan a hardware purchase around specs that do not exist.

For now, if you were weighing local hardware against the API cost of MiniMax H3, the choice is made for you: the API is the only route. Re-check MiniMax’s official channels for the weight release and its stated GPU requirements before budgeting for a local setup.