+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Security, Privacy & Safety (https://www.datasculpting.ai/inference/forumdisplay.php?fid=23)
+--- Forum: Model / Agent Security (https://www.datasculpting.ai/inference/forumdisplay.php?fid=77)
+--- Thread: tool-calling security concerns: worth it? (/showthread.php?tid=50)
tool-calling security concerns: worth it? - benchplease - 08-28-2026
Did a bit of testing and thought I'd post the result here.
I tried this with Mistral using LM Studio on GB200. The headline result was better than I expected, but VRAM headroom still ended up being the limiting factor.
Happy to post more details if anyone wants them.
Re: tool-calling security concerns: worth it? - rasterhall - 08-29-2026
What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?
Re: tool-calling security concerns: worth it? - lablab - 08-29-2026
Did a small comparison this afternoon. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.