LFM3 2B Omni
on GTX 1050 4GB · Ollama · 32,768 ctx
Use cases
vision
Summary
User reports LFM3 Omni 2B at 43.7 tok/s on a laptop GTX 1050 4GB.
Setup is Ollama with a native 32K context, using 2,479 MiB of VRAM.
User says the model is very fast and its image features are good, but its text responses are poor. Other models tried (Qwen 3.5, Qwen Coder, Nemotron 3 Nano, Bonsai 2, LFM2.5) are named without figures.