Edge Compute

Liquid AI released LFM3 models sharing a backbone across text, vision, audio and decision tasks. Ollama says the 1.7B text model supports 256K context and fits in about 1 GB at 4-bit precision.

Perplexity Computer adds local model and agent execution on Macs using the Qwen-derived PPLX Pro model. Local tasks use no Computer credits, while 36 GB of unified memory is recommended.

Vik Paruchuri released Indra, a 1M-parameter OCR model with code and a local viewer. He reports higher OmniParseBench text scores than PaddleOCR and Tesseract, but notes limited handwriting support and no multi-page context.

OpenBMB released open MiniCPM-V 4.7 models in 35B-A3B and 1B sizes with mobile deployment support. It reports fivefold faster inference on a 16GB iPad Pro M4 using Lookahead Expert Scheduling.

Google released EmbeddingGemma 2 under Apache 2.0 to map text, code, images, video, and audio into one embedding space. The 740M-parameter version supports 8K context and on-device retrieval.

Mario Zechner demonstrates a native Android app built on his phone with Pi Durable in two days. The prototype also adds paged transcript loading, but he describes it as a demo rather than a product.

Mario Zechner's Pi Durable prototype runs an agent runtime on Android while calling remote models. Shared sessions and subagents are supported, and work resumes after worker restarts.

Meta's Muse Gadgets provides open-source ESP32 firmware and a Linux SDK for connected hardware. Meta also announced Muse Home Link for compatible smart devices, offering 5,000 units free to eligible US Muse subscribers.

Black Forest Labs released FLUX 3 Action, an open 7B model that jointly predicts future video and actions for robot policies. The company reports first place on RoboLab and released embodiment fine-tunes, training recipes, and Jetson deployment support.

Perplexity launched Portable Computer in its Windows app for supported NVIDIA RTX systems. Local inference requires at least 24GB of VRAM, while the update also adds local MCP support and scheduled tasks.

Perplexity open-sourced Lily, a Rust and Metal engine for Qwen3.6-35B-A3B in Perplexity Computer's hybrid workflow. Perplexity reports 1.23× faster prefill and 1.35× faster decode on an M5 Max MacBook Pro.

Perplexity's Mac app can route sensitive agent steps to local models while using cloud models for other work. The company also open-sourced the PII classifier used to decide where work runs.

A developer reports streaming 60% of Qwen3.8-Flash-Next experts from disk on demand, running the full model in 37 GB at 40 tokens per second. BF16 and GGUF weights are also available for local deployments.

Perplexity’s Portable Computer runs its orchestrator, subagents, and harness locally on NVIDIA DGX Spark with a post-trained 27B model. Frontier-model escalation requires user approval and flags PII before text is sent externally.

A LocalLLaMA post ran a 122B Qwen model on an older laptop with llama.cpp. The run had very long load and generation times, while another report put Qwen 3.6 35B at 21 tok/s on a Radeon 7600 after ROCm tuning.

Magnitude launched an open-source terminal coding agent that runs local models on-device without API keys. Its launch post says it profiles hardware and can use shell, file-editing, script, and skills tools.