EstateDraftFL adds a three-engine AI roster: Claude Opus 5, ChatGPT 5.6 Sol, and Claude Fable 5

Editor's note (2026-09-05)

Since UPL Program v3 (2026-09-03) these engines run only inside law-firm workspaces. EstateDraftFL's self-help documents are assembled deterministically from the customer's own answers; no model writes, phrases, or reviews them. The release below is kept as published.

One cage, three engines

Every engine runs behind the same gateway: the same spend governance, the same audit trail, and the same operative-text validation cage that keeps AI output in an informing role. The engines differ in reasoning character and cost; the guardrails do not.

Engine choice is available on the full versions of EstateDraftFL's AI reports — Estate Plan Review, Opposing Counsel Pass, Redline, and Cite-Check — and in the drafting workstation Florida law firms use with their own supervising attorneys. Free previews continue to run on Claude Opus 5.

What each engine is for

Claude Opus 5 is the house engine: the default for every AI feature on the platform. ChatGPT 5.6 Sol is a peer-tier alternative for teams that want a second frontier perspective on the same material. Claude Fable 5 is the premium option — a heavier model offered where the extra depth is worth the extra cost.

Consumer document drafting itself remains deterministic: the documents a member assembles are built from attorney-authored Florida templates driven by the member's own answers, not generated by a language model. The engines read, review, and explain; the document engine assembles.

For Florida firms

Firm workstations add an attorney-gated drafting lane where a supervising attorney chooses the engine per matter and reviews everything before it moves. Firm tenants keep their included AI allowance, itemized engine pricing, and the same tenant-isolated audit trail.