Observed arrival · 2026-10-25
A VRAM budget for Skyrim’s AI-NPC stack
A calculator estimates whether a GPU can run a selected game, AI NPC mod, local language model, and voice setup together—or whether cloud inference may be a better fit.
- For
- Skyrim players configuring local AI NPC mods
- Worth noticing
- The estimate itemizes game memory, model weights, KV cache, runtime, voice, speech recognition, and OS reserve.
Field notes
Inputs extend beyond a simple GPU lookup: users can choose texture presets, model quantization and context length, then add separate text-to-speech and speech-recognition workloads. A sample 12 GB setup totals 11.2 GB and is marked tight, with suggestions including lowering context or moving the language model to a cloud backend. The page attributes its method to published information from Mantella, CHIM, SkyrimNet, and model files on Hugging Face.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue