inferd
John Davies launches inferd, an open source Mac inference engine for 64 to 128 concurrent agents
Launched live on ThursdAI by Incept5 CTO John Davies: an Apache 2.0 inference engine for Macs (laptops too) built for many agents at once. It went from 64 concurrent agents to 128 on the day of the show, and on a 32 GB Mac Mini it runs Qwen 3.8 27B with a DFlash 2 drafter at up to 16 agents. In John's chart, other engines flatline around 400 tokens per second at 16 agents while inferd keeps climbing past 1,000 tokens per second in total.