LiquidAI's LFM3 runs on your phone; Nikita Bier's local-model thesis resurfaces
maximelabonne · x · 2026-11-22
@helloiamleonie shared LiquidAI's LFM3 family — text, vision, audio, omni, pick what you need and put it on your phone, with GGUF/ONNX/MLX/FP8/NVFP4 quantizations available. The quote from Nikita Bier argues a massive market opportunity would be unlocked by good local mobile models, since phones hold data too private and too large to send to frontier labs.
On HF, LFM3-1.7B has 30k+ downloads and LFM3-VL-2B (1.9B params, image-text-to-text) has 288 likes.
More from Infra
- WSJ: Associate of Jensen Huang's Partner Accused of Funneling Nvidia Chips to China — GaryMarcus · 2026-11-22
- Hierarchos-Native v0.2.5 ships P3 FP32-merge for RDNA3 WMMA GEMM, up to 27% faster — PhysicsDisastrous462 · 2026-11-22
- prime-rl 0.11.0 ships: 1.5x faster training, new pretraining entrypoint, Mega MoE dispatcher — samsja19 · 2026-11-22
- Math suggests 7.7M GPUs would serve everyone: fuel for the AI overbuild hypothesis — bookwormengr · 2026-11-22
- Chinese NAND maker YMTC promotes ZHITAI SSDs at Korea's G-Star game show — terryyuezhuo · 2026-11-22
- SGLang runs Inkling 2 Mini in 65GB on DGX Spark, 2x faster prefill than vLLM — AIFlow_ML · 2026-11-22