한국어EN
On-Device & Local InferenceGoogle (AI Edge)Google

Google makes LiteRT-LM generally available (GA)

LiteRT-LM generally available (GA) — added a native Swift API for iOS and a WebGPU-accelerated JavaScript API. Supports Android CPU, OpenCL GPU and NPU, iOS Metal and browser WebGPU. With Gemma 4: 52 tok/s decode on Android GPU, 56 tok/s on iOS GPU, 76 tok/s on the web on a MacBook Pro M4 Max, up to 2.2× with Multi-Token Prediction, 607MB of memory on iOS

Why it matters

The stage at which an on-device runtime covers not just mobile OSes but browsers with the same stack. Figures are the vendor's own measurements

Source — Primary sources first; secondary reports are labelled “Press”

Officialdevelopers.googleblog.comhttps://developers.googleblog.com/blazing-fast-on-device-genai-with-litert-lm/

Added in Week 36, 2026 · See that week

Same week

Contextual and parametric faithfulness shown to correlate only asymmetricallyPaperAnthropic researchers publish Faithfulness as Information FlowPaperScience ministry pre-announces amendments to the AI Basic Act enforcement decreePressCohere releases Command A+Official

Nearby on the same axis

llama.cpp Discussion #23853 shares an NVFP4/MXFP6 mixed-precision toolOfficialChrome 148 stable opens the Prompt API to the webOfficialOllama 0.30 adds llama.cpp-based GGUF compatibility to its MLX engineOfficialApple expands the Foundation Models framework at WWDC 2026Press