Small / on-device LLMs
Small / on-device LLMs covers models built to run on phones, laptops, browsers and edge devices (roughly under 4B params).
In brief
Small / on-device LLMs, in brief.
What it isSmall / on-device LLMs covers models built to run on phones, laptops, browsers and edge devices (roughly under 4B params).
Where it sitsOne of 24 categories in the Models layer. Models are the trained systems that do the work: language models, and models for speech, images, video, code, translation and forecasting. This layer covers both closed models reached through an API and open models you can download.
What we track24 products, including Apple Foundation Models (on-device ~3B), DeepSeek-R1-Distill-Qwen-1.5B, EXAONE 4.0 1.2B, Falcon-H1 (0.5B-3B) and Gemini Nano (Android AICore / Chrome built-in AI).
The products
24 products in Small / on-device LLMs.
Each name opens a search for that product in our index.
ProductMakerIts own site
From our census of 2026-11-08. A product listed here is one that belongs in this category; the search answer says whether we have checked it.
Questions
Small / on-device LLMs, answered.
What is Small / on-device LLMs in AI infrastructure?Small / on-device LLMs covers models built to run on phones, laptops, browsers and edge devices (roughly under 4B params).
Which layer of AI infrastructure is Small / on-device LLMs in?Small / on-device LLMs is in the Models layer. Models are the trained systems that do the work: language models, and models for speech, images, video, code, translation and forecasting. This layer covers both closed models reached through an API and open models you can download.
Which products are in Small / on-device LLMs?We track 24 products in Small / on-device LLMs, including Apple Foundation Models (on-device ~3B), DeepSeek-R1-Distill-Qwen-1.5B, EXAONE 4.0 1.2B, Falcon-H1 (0.5B-3B), Gemini Nano (Android AICore / Chrome built-in AI), Gemma 3 / Gemma 3n (270M-4B, E2B/E4B), Granite 4.0 Micro / Nano and H2O-Danube3. The full list is on this page.
Models
The other categories in Models.
3D generation & motionAvatar / digital humansBackground removal & mattingBenchmarks & leaderboardsCode modelsEmbedding modelsFrontier model APIsImage generation & editingImage upscaling & restorationLip-sync & dubbingMachine translation APIsModel hubs / registriesMultilingual / regional-language modelsMultimodal / vision-language modelsMusic & audio generationOpen-weight model familiesRealtime speech-to-speech modelsReasoning modelsRerankersSpeech (STT/TTS)Time-series foundation modelsVideo generationWorld models