CoreAI Zoo hits App Store: run Qwen, Gemma and more fully on-device on iPhone
pcuenq · x · 2026-09-19
- CoreAI Zoo is now on the App Store (free, BSD-3-Clause open source), built on the new CoreAI on-device framework. It runs open-weight models entirely locally on iPhone, iPad and Apple silicon Macs — no account, no server, nothing leaves the device.
- Coverage: chat (Qwen3, Gemma, LFM2.5, Granite, Mistral, 0.6B to 30B+ depending on memory); vision Q&A and OCR (Qwen3-VL, MiniCPM-V); transcription (Whisper large-v3-turbo, Parakeet) plus speech and music generation; live camera depth estimation, detection (YOLOX) and super-resolution; experimental diffusion LMs.
- How it works: everything runs via Core ML on Apple silicon; weights download once from the Hugging Face Hub and are cached for offline use; one model resident at a time; new models arrive via a live catalog without app updates. A BENCH tab measures load time, tokens/s and memory.
- Getting started: begin with Qwen3-0.6B (454 MB); recent iPhone Pros or Apple silicon Macs handle mid-size models comfortably.
More from Infra
- Jev scales on Modal, one of only 4 subprocessors listed, to meet surging demand — AAAzzam · 2026-09-19
- Tuning Qwen3 27B as a coding agent on 2x3090s cuts turn latency from 28s to 7s — bolts98 · 2026-09-19
- AgentZip: memory compression for parallel agent sandboxes cuts memory up to 8.7x — rohanpaul_ai · 2026-09-19
- US products quietly build on Chinese open-weight models as one firm cuts spend by ~100x — generativist · 2026-09-19
- Turbines sold out through 2030: 25 categories of datacenter power gear face multi-year backlogs — CatAstro_Piyush · 2026-09-19
- Why OpenAI and Anthropic are suddenly buying tiny 20-30MW data centers — abhiadesai · 2026-09-19