Liquid AI's LFM3 Is CRAZY: 1.7B Model vs 9B? I Tested It on the Live API

Liquid AI's new LFM3 family: six ~2B open-weight models (text, vision, audio, omni, decision). I walk through all of them, test four on the live API, show where one failed, and say where bigger models still win. Benchmark numbers are Liquid AI's own (vendor claims) from liquid.ai/blog/lfm3; the API tests are mine. Chapters: 0:00 Cold open: 1.7B vs 9B 0:53 What LFM3 is 1:42 Architecture: 13 conv, 5 SWA, 6 global 3:09 Text model LFM3-1.7B 5:27 Live API test: text + tool calling 6:05 Vision LFM3-VL-2B (and where it failed) 7:44 Audio LFM3-Audio-2B 8:47 Omni LFM3-Omni-2B 10:11 Decision models d1 11:32 Fine-tuning and robotics 12:34 Verdict, limits, where to run it Links: Blog: https://www.liquid.ai/blog/lfm3 Docs: https://docs.liquid.ai/lfm/models/complete-library Hugging Face: https://huggingface.co/collections/LiquidAI/lfm3 Playground: https://playground.liquid.ai API console: https://console.liquid.ai Cheap GPUs: RunPod: https://get.runpod.io/pe48 Credits: short demo clips shown are Liquid AI's own official demos, labelled on screen. Voice is my AI-cloned voice. Thumbnail is a composite of my photo and a real page crop. Get GPUs Runpod: https://get.runpod.io/pe48 Get Hostinger: https://www.hostg.xyz/SHJiu Coupon Code "PROMPT" 10% discount 🔗 Connect with me: Email → [email protected] Ko-fi → https://ko-fi.com/promptengineer Patreon → https://www.patreon.com/PromptEngineer975 Book a Call → https://calendly.com/prompt-engineer48/call GitHub → https://github.com/PromptEngineer48 Twitter/X → https://twitter.com/prompt48 Website → https://promptengineer48.com/