Reporter Tests M5 Ultra Local AI: Hermes Paired with Qwen Models
Understand the story
On October 11, 2026, a writer at The Verge began hands-on testing of running large language models locally on an M5 Ultra Mac Studio. The test machine has 256GB of unified memory. The writer installed Hermes, an open-source self-hosted Agent desktop app, and chose the Qwen 3.8 Flash Next model, which has 125B parameters and a size of about 105GB. To make it easier to operate from a phone, the writer also set up remote control of the local AI via a Telegram bot. The test's focus scenarios are using the Hermes Agent to organize a Steam game library and handle private data. The hands-on report is still in its early stages, and no further test results or conclusions have been disclosed yet.
AI overview of reporting · Updated 7 days ago
Reporting timeline
Explore different aspects of the story through its reporting.
- The Verge · AILearning to use local AI is exciting, overwhelming, and frustrating
A Verge writer began testing running LLMs locally on an M5 Ultra Mac Studio with 256GB unified memory, installing the open-source self-hosted agent desktop app Hermes, choosing the 125B-parameter, roughly 105GB Qwen 3.8 Flash Next model, and adding control from a phone via a Telegram bot.
Story heat trend
Current heat3·Comparable peak9（10/11 21:00）·Comparable change over 24 hours -49%
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.