Raffaele Mancini

[Raffaele Mancini @RaffoMancio]

@ivanfioravanti Don't worry; if you have an old Android phone, you can run Qwen 3.8 Flash Next on it too

[Raffaele Mancini @RaffoMancio]

@MiaAI_lab It fits on a android 12gb ram mobile phone

[Raffaele Mancini @RaffoMancio]

@0xBakeer https://t.co/O3K1x3ksVc

[Raffaele Mancini @RaffoMancio]

@TeksEdge https://t.co/O3K1x3ksVc

[Raffaele Mancini @RaffoMancio]

@TeksEdge Imagine running Qwen 3.8 Flash Next on a 12GB Android phone

[Raffaele Mancini @RaffoMancio]

@0xSero And phone.
Imagine running Qwen 3.8 Flash Next on a 12GB Android phone - 2 tok/s

[Raffaele Mancini @RaffoMancio]

@0xBakeer @Alibaba_Qwen @QwenDevs Imagine running Qwen 3.8 Flash Next on a 12GB Android phone

[Raffaele Mancini @RaffoMancio]

@mudler_it Imagine running Qwen 3.8 Flash Next on a 12GB Android phone

[Raffaele Mancini @RaffoMancio]

@OpenRouter @Alibaba_Qwen Running Qwen 3.8 Flash Next on a 12GB Android phone
https://t.co/O3K1x3ksVc

[Raffaele Mancini @RaffoMancio]

@UnslothAI Running Qwen 3.8 Flash Next on a 12GB Android phone with unsloth gguf. 2 tok/s
https://t.co/O3K1x3ksVc

[Raffaele Mancini @RaffoMancio]

@shuai_bai_ Running Qwen 3.8 Flash Next on a 12GB Android phone - 2 tok/s
https://t.co/O3K1x3ksVc

[Raffaele Mancini @RaffoMancio]

@Alibaba_Qwen https://t.co/O3K1x3l0KK

[Raffaele Mancini @RaffoMancio]

@Alibaba_Qwen 3.8 Flash Next runs perfectly on my Android phone (12GB RAM)

As you know, my Bigmoeonedge project enables running massive models on edge devices - such as a mid-range Android phone with 12GB of RAM.

Following DeepSeek and various other models, Qwen 3.8 Flash Next is now supported too, just hours after its launch via a PR and a dedicated branch in the open-source project.

The demo shows generation speeds of around 2 tokens/s.

Qwen's new architecture is perfect for this use case, and I’m happy with how easily I was able to integrate it into the project.

A big thank you to @ggml_org , my contributors, and everyone supporting the project! Feedback is always welcome!

P.S. This is just the beginning; I plan to boost performance within a few weeks. Stay tuned.

[Arduino @arduino]

The wait is over! Meet Arduino VENTUNO Q, where AI takes action.

💬 Run local LLMs like Qwen 3, Gemma 4, and Qwen 3 VLM directly on the board
🧠 NPU + CPU + GPU +MCU: @Qualcomm Dragonwing IQ-8275 with up to 40 dense TOPS of AI performance and STM32H5 microcontroller for real-time control
🗄️ 16 GB RAM + 64 GB eMMC + expandable storage
🪁 Linux-powered, pre-loaded with Ubuntu OS + Zephyr RTOS
🛠️ Build AI faster using Arduino App Lab, @huggingface, @EdgeImpulse, and Qualcomm AI Hub
💨 Move seamlessly from prototype to production through Works with Arduino Certification Program

The first batch won’t last. Pre-order your VENTUNO Q with a free power supply and USB-C cable included! Get ready to enter a new era of Physical AI: https://t.co/YpDhKRuX6y

[Raffaele Mancini @RaffoMancio]

👀👀👀

[ModelScope @ModelScope2022]

The next-gen architecture powering Qwen4 is now here! ✨
Get ready for the open release of Qwen3.8-Flash-Next 🚀
The countdown starts now! ⏳🔥https://t.co/0Zlesgxnot

[Raffaele Mancini @RaffoMancio]

If you want to try it on your android phone, BigMoeOnEdge supports Ling-3.0-flash and tiny, hours after llama.cpp merged the architecture upstream.

The experts stream from flash on demand, so the whole thing runs on a android 12 GB phone.
Built with llama.cpp, one registry row.

2.6 tok/s flash
16 tok/s tiny

[Raffaele Mancini @RaffoMancio]

BigMoeOnEdge support for @AntLingAGI 's Ling-3.0-flash and tiny, hours after @ggml_org llama.cpp merged the architecture upstream.
124B, 5B active.
The experts stream from flash on demand, so the whole thing runs on a 12 GB phone: 51 GB on disk at 3-bit, four times the RAM.
Built with llama.cpp, one registry row.
2.6 tok/s flash
16 tok/s tiny