Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #7960

LFM2.5-VL-3B

Related event timeline, sources and context from the news index.

EVENT TIMELINE

2

MODELS · 1 SOURCE · Hugging Face

Open multimodal decision models d1-3B and d1-omni-600M released for edge inference

The team released two open-weight decision models, d1-3B and experimental d1-omni-600M, built on their Liquid Foundation Models. d1-3B (decoder-only) supports text+images, scores 48.57 on Decision Index v0.2.1 (best under 10B) and runs in 16–50 ms on NVIDIA Jetson devices; d1-omni-600M (encoder-based) supports text+image or text+audio, is an early research release, and both models (and demos) are available on Hugging Face.

8.0

MODELS · 1 SOURCE · Hugging Face

Liquid AI publishes DSpark drafter for LFM2.5-VL-3B, accelerating VLM decoding up to 3.13×

Liquid AI released a DSpark draft model for its vision‑language model LFM2.5‑VL‑3B that adds speculative decoding to speed up token decoding (up to 3.13× on-device, 2.66× on H100) while increasing model size by ~280M parameters (≈8.9%). The drafter ships with day‑one integrations for llama.cpp, MLX‑VLM and SGLang and is available on Hugging Face in Safetensors and GGUF formats; measured end‑to‑end gains ranged up to 2.62× on edge and 2.27× on GPU across MMSpec tasks.

7.0