>>

Run granite-switch-4.1-8b on Lium

IBM · Text generation · 8.8B parameters · 17.6 GB BF16 weights · released 2026-10-14
It needs about 22 GB of VRAM. The cheapest setup that runs it is RTX 3090 (24 GB), from $0.19 per hour, billed per second.
Rent RTX 3090

Other GPUs that run it

RTX PRO 4000 Blackwell24 GBfrom $0.40 / hourRent
2x Tesla V10032 GBfrom $0.42 / hourRent

How to run it

Rent the pod above, open a terminal over SSH or Jupyter and serve ibm-granite/granite-switch-4.1-8b:
vllm serve ibm-granite/granite-switch-4.1-8b
Claude Code and Codex can do all of it through the Lium CLI.

Frequently asked questions

How much VRAM does granite-switch-4.1-8b need?

About 17.6 GB for the released BF16 weights, so plan on 22 GB with room for activations and KV cache at a short context. Longer contexts and bigger batches need more.

What GPU should I rent to run granite-switch-4.1-8b?

RTX 3090 (24 GB). On Lium that setup costs from $0.19 per hour, billed per second with no minimum.
All open models
For AI agents (CLI install, agent skill, headless signup): llms-full.txt