Ben Buschmann

We added support for Imagination Studio by Keys + Wasserman first (@SamJWasserman @u1tra_instinct). GapRunner now schedules GPU work it didn't start: an app asks for a reservation, waits its turn, and GapRunner makes room. The LLM sleeps; training parks at its last saved step.

I open-sourced GapRunner: a job queue for the GPU that runs your local model.

Training runs, evals and benchmarks wait their turn and run beside the model if they fit, in its idle gaps if not. A request comes in, they step aside, the model wakes.

github.com/benbuschmann/g…

1/
Benchmarked Apple's Private Cloud Compute model (AFM 3 Cloud) on macOS 27.

fm serve gives you an OpenAI-style endpoint on localhost, so this was curl and a Python timer.

Prompt processing: ~2,300 tok/s
Generation: 118–143 tok/s
A 26K-token essay summarized in 15s

Early Darkbloom benchmarks (qwen3.6-35b-a3b-vl-mtp-mxfp8, decode tok/s):

🚀 Base M5 Max Studio: 122
💡 M6 Mini: 70
The M6 Mini beats my M1 Max (65) with less than half the memory bandwidth. The M5 Max is on another level.

anyone see benchmarks for the new m6 mini running local ai models? if so please post. The reviews I see just talk about Geekbench. I think m6 is a sleeper. I have a 32gb version coming tomorrow so I will be testing.