Skip to content

Is it a fit?Book a call

ManifestoEvery company, and one day every home, will run its own AI.Read it in two minutes.

Owned AI overview Eigna and the alternatives See if it’s a fit

By company size

Finance and law

Public and critical

Industry

Creative and digital

Industries overview All industries See if it’s a fit

Book a call See if it’s a fit

Insights

What we released, what we held, and why.

News, explainers, research, tests and cases. Every number with its baseline and its date.

Seven slits of light cast across graphite

FeaturedJournal26 October 20265 min

Benchmark 001, losses included

Why we published our first benchmark in full, including the line where we are not yet ahead.

Omar GarciaRead the piece

October 202614 pieces

A row of standing plates, each lit by the Figure, receding into the dark

Journal26 October 20263 minRollback in 60 secondsYour model stays the same until you decide. If a release ever disappoints, the previous one is one step back. Why we design the way back first.

Eigna Box Home standing upright, the Figure cut through its face

Journal26 October 20263 minInside the Eigna BoxOwning AI is easier to believe when you can point at it. One base, five homes for the model, and why we build the first ones ourselves.

An aluminium key with the Figure cut through it, standing in soft light

News26 October 20261 minEigna Labs, Inc. is formedEigna Labs, Inc. was formed in Delaware in October 2026, to build AI that belongs to the company that uses it.

The Figure cast as light on mineral green plaster

Research26 October 20263 minLearning on small approved dataA model of your own learns only from the work your people approve. That is very little data. Four open problems, and what we know so far.

Five small figures etched in a row on brushed aluminium, one per release

Journal25 October 20263 minHow a release earns its placeA new version of your model goes live only when it beats the one you use today. This is the test it has to pass.

The Figure cast as light on blue-grey plaster

Journal25 October 20263 minWhy there is no token billRented AI gets cheaper per token and more expensive per month. An owned model runs at a fixed cost. Where that wins, and where it does not.

The Figure as seven crossings of light on graphite

Journal25 October 20263 minWhy our mark is made of lightTwo plates, seven crossings, and light only where both are open. The Eigna mark is the whole idea in one shape.

Seven crossings cut through a dark plate, lit from behind

News25 October 20261 minInterference: the Eigna identityOur brand guidelines, version 1.0: two plates, seven crossings, and the Figure that appears only with your work.

A field of fine slits forming the Figure in light

Research25 October 20263 minThe Proof Ladder: what we test nextSix rungs, nine model families, 54 cells. One is done. What each rung proves, and what runs next.

The Figure cast as light on a mineral green sphere

Case25 October 20268 minCase 001: held at the gateA forecasting model of its own. Measured in the open, held at the gate.

Test24 October 20263 minThe Proof LadderSix rungs, nine model families, 54 cells. One done. What runs next.

A hand writing on cream paper on dark slate, the Figure drawn at the top of the sheet

Test24 October 20263 minMethods and correctionsMetrics, test design, digests to reproduce, limits, and the standing corrections rule.

The Figure cast as light on graphite plaster

News14 October 20261 minBenchmark 001 is published, losses includedOur first training run cut forecast error 17.8% against its base model. It is not yet ahead of a simple last value baseline, and we publish that too.

1.750%1.439%1.392%Test14 October 20267 minBenchmark 00117.8% lower forecast error than its base model. Not yet ahead of a last value baseline.