BUILD JOURNAL / PUBLIC RECORD

We will show
the work.

BrokenGPT is being built in three phases. This journal will publish the methods, evaluations, costs, failures, and model decisions behind each one—not just the launch announcement.

DISPATCH 001PHASE 01 / LIVE

JULY 2026 · FOUNDATION NOTE

Why we are starting with disclosed model access.

Phase 1 gives us a measurable baseline. Disclosed third-party models power chat and the compatible API while the same gateway records usage, latency, and product behavior. Starting here lets us separate improvements to the model from improvements to the service around it.

The result is a working foundation and a record we can compare against when Phase 2 begins. Model identity, context limits, metering, and the service boundary remain visible throughout the process.

See the complete roadmap

DISPATCH 002 · TARGET OCT 2026

From refusal reduction to full-precision evaluation.

How we select checkpoints, build refusal and accuracy evaluations, test regressions, and decide whether a change is actually better.

PHASE 02 REPORT / PLANNED

DISPATCH 003 · TARGET EARLY 2027

Training a BrokenGPT-native model.

The data decisions, training approach, capability results, limitations, and deployment architecture behind our own uncensored model.

PHASE 03 REPORT / PLANNED

OUR EDITORIAL CONTRACT

No victory lap without the receipts.

Methods

What we changed, why we changed it, and the tools used to get there.

Measurements

Capability, refusal, accuracy, and regression evaluations—not cherry-picked demos.

Tradeoffs

The cost, latency, serving precision, and behavior compromises behind each release.

Failures

Experiments that did not work and the evidence that made us change direction.

PHASE 01 IS LIVE

Try the foundation.

Use the product now, then follow the journal as the work moves from disclosed model access to original research.

Open chat