Get Zlio on your home screenOne tap, full screen, like an app.
Developer Tools · first seen 05 November 2026, 18:23 UTC
Delta Inference Engine v0.93 adds multi-GPU, faster first token
The Delta Inference Engine, a local tool for running large models, hit v0.93: DeepSeek-V4.1-Flash on one RTX 3090 at 41 tok/s, support for 64GB RAM, Unsloth quants, earlier first token and basic multi-GPU.