Every Orion training run, with its numbers, dates and sources. Where the published record is silent, the page says unknown.
Three steps between the first launch and the 100B run.
The launch run, permissionless. It "fell over almost immediately" under system load and adversarial attacks.
A 1.5B, 3-layer testbed; almost 15 trillion tokens in total; system throughput up about tenfold.
Model size up 67x and pipeline stages up 10x in a single month. More than 750 experiments in all, from 1.5B to 21B, before Orion-100B.
Exploit Summit, Montreal, 28 September 2026. "Project Orion: Introducing Liquid Training," Will Squires and Steffen Cruz. A stage figure enters the register only when Macrocosmos publishes it as a run.
"Powers any training workload across it, as if it were one cluster."
The iota SDK and Liquid Compute, "our disaggregated compute platform." Go-to-market "in the coming weeks"; early access by registration.
Macrocosmos on X, about 15:30 Montreal time."Just a few lines changed from pure PyTorch."
Two years of research "distilled into a powerful set of communication primitives."
Steffen Cruz on X, the same hour."3x cheaper than if we had simply reserved a node in a data center."
The 1 June write-up puts entry cost per replica 2.5x below a Covenant-class 8×B200 peer; the 25 September post says the same.
Steffen Cruz, quoted by The TAO Daily with a 3:21 clip.164 companies "already in conversation." Pilots "from October."
100 billion tokens trained; a 16B model at 24% MFU; Macrocosmos's 23 September post gave about 20% for the first run. A path "through ResBM and ResBMoE toward 1T and 10T models."
Subnet Summer, live from the stage."All of our published results to date are on the legacy code base."
Pretraining "is too small of a market," so the focus now covers RL, fine-tuning and more.
Will Squires on X, 28 September.Replay: a 38-second highlight on the summit's stream siteverified.