Building on the latest models — with Copilot joining the coders
AIFactory builds software autonomously: it takes a plan, writes the code in an isolated git worktree, and opens a pull request. This round of work moved it onto the current models, added a third coding runtime, and connected it to the rest of the family.

On the latest models
The default coder is now Claude Opus 4.8, Anthropic's current flagship. Claude Sonnet 5 and the current OpenAI Codex, Google Gemini, and GitHub Copilot models are all selectable per stage. We refreshed the whole model catalogue in one pass — and fixed a stale price entry that had been over-stating Opus cost three-fold.
Three coding runtimes
Claude, OpenAI Codex, and GitHub Copilot now all run as first-class coding backends behind one task contract. The Copilot CLI is baked into the build image and wired as a selectable runtime; turning it fully on is one operator sign-in away.
Part of one product
AIFactory also joined the fleet-wide unified shell:
- A portal switcher moves you between Plan, Build, Test, and Cockpit.
- A global command palette — Cmd-K — searches every portal's work at once.
- A fleet "needs you" badge shows how many tasks are waiting on a human.
- Silent single sign-on carries you in from any sibling portal.
Download the showcase
AIFactory — one-page showcase (PDF)
The path forward
Next: bring the GitHub Copilot runtime fully live, sharpen acceptance-criteria fidelity, and keep the build stage able to ship any language to any target.
