About

Frontier AI
doesn't need a
datacenter.

Ternative is an open-source AI project from Colombia, building competitive language models — and the inference engine to run them — on hardware ordinary people already own.


The thesis

Constraints as
a method.

The dominant story of modern AI is one of scale: more parameters, more GPUs, more capital. Terse is our argument against that being the only path.

Terse is a family of ternary-weight models built from scratch — our own architecture, every internal weight −1, 0 or +1. Terse Micro, the first, folds a whole language model into a single ~182 MB file that runs on a plain CPU. It isn't a frontier model, and we don't pretend it is — but building a coherent ternary model from the ground up, on a weekend's worth of compute, is the harder and more durable breakthrough: the footprint win compounds as the family scales.

It grew out of Orchid 1.0, our earlier 2B model — a fine-tune of Microsoft's BitNet b1.58, aligned with ORPO on a single 4 GB laptop, and the first competitive LLM trained and aligned in Colombia. Orchid proved the idea; Terse rebuilds it from first principles. Serving either one meant solving problems the big labs never hit, so we built the ternative engine — all of it Apache 2.0, with weights, code, and an archived DOI.


One project, three names

Ternative, Terse, Orchid.

{ }

ternative

The project itself — and its in-house x86/AVX2 inference engine for ternary-weight models.

−10+1

terse

The from-scratch family of ternary-weight models — their own architecture, starting with Terse Micro.

Orchid pixel-flower logo

orchid

Orchid 1.0, our earlier model — a fine-tune of Microsoft BitNet b1.58, not an own architecture.


What we stand for

Reproducible, local, open.

−10+1

Reproducible

A documented recipe with published failure modes — not a black box. Reproduce the benchmarks with the scripts in the repo.

local-first

Local-first

Inference runs on your machine — CPU-only if needed. No account, no telemetry, no cloud dependency.

Apache 2.0

Open

Weights, engine, paper and data recipe — all public, free for research and commercial use alike.


Sponsor the work

Keep independent AI
free.

Ternative is built and maintained in the open, outside any large lab. Funding goes directly to continued development — better models and a faster engine. If your organization relies on open, reproducible AI, consider supporting it through FLOSS/fund.

Apache 2.0
Permissive license — yours to build on
100% open
Weights · engine · paper · data recipe

Everything, in the open

Project links

Hugging Face
Model card, GGUF weights & LoRA adapter
huggingface.co ↗
ternative engine
C++17 / CUDA inference engine — source & releases
github.com ↗
orchid-1.0 recipe
Training recipe, eval harness & reproduction scripts
github.com ↗
Zenodo · Technical paper
DOI 10.5281/zenodo.20452163 — archived research record
zenodo.org ↗
FLOSS/fund
Fund continued open development
floss.fund ↗

Cite the work

Used Orchid or ternative
in research?

Both have a citable record. The model is archived on Zenodo with a permanent DOI.

citation.bib
@software{romerochisco2026ternative,
  title  = {ternative: Inference Engine for
            Ternary-Weight LLMs with Runtime LoRA},
  author = {Romero Chisco, Michelangelo},
  year   = {2026},
  license = {Apache-2.0}
}