Liquid AI Adds Speculative Decoding to LFM2.5-VL-3B for Faster Vision-Language Inference
Liquid AI released LFM2.5-VL-DSpark, a 280M-parameter drafter that speeds LFM2.5-VL-3B decoding on edge devices and H100 GPUs.
Latest News and Analysis in On-Device AI
Liquid AI released LFM2.5-VL-DSpark, a 280M-parameter drafter that speeds LFM2.5-VL-3B decoding on edge devices and H100 GPUs.
PrismML released Bonsai 2 27B, a compressed reasoning model designed for PCs and phones, strengthening the case for private, on-device AI.
Google says Gemma 4 has passed 150 million downloads and showcased builders using the open models for on-device AI apps.
Google introduced Gemma 4 12B, an encoder-free multimodal open model built to run locally on laptops with 16GB of memory.
Reports say Chrome has been installing a local Gemini Nano model for AI features, raising storage, consent, and energy concerns.
Google released an offline-first AI dictation app for iOS powered by its Gemma AI models, taking aim at competitors like Wispr Flow.
As Apple celebrates its 50th anniversary, former insiders say the company squandered an early AI advantage but is now betting on on-device AI and Google Gemini to reboot Siri.