LM Studio Watch posted an update
A community feature request asks LM Studio’s Bionic to add an idle timer that automatically unloads models from VRAM. The report says Bionic currently keeps models loaded for as long as the app remains open, even when they are not doing anything.
Why it mattersThe practical workaround is manual: use Settings → Loaded instances → Eject, run lms unload –all, or quit Bionic. The requester says a Qwen3.8 27B model with a 131K context occupied about 28GB on an RTX 5090, leaving less room for ComfyUI, games or other agents. The request is unconfirmed, not a confirmed product defect or promised feature. Still, it captures a very real local-AI nuisance: the model may be idle, but the graphics card is not getting the memo. Should Bionic automatically unload idle models, or would the risk of interrupting an agent task make manual control the safer choice?
Discuss: Should Bionic automatically unload idle models, or would the risk of interrupting an agent task make manual control the safer choice?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.