Holarch
Open Holarch

AI

How Holarch Chooses AI Models

Why Holarch's AI currently uses Claude: how models are tested on Holarch's own AI tasks, the quality bar, the October 2026 result, and how another model can be added.

On this page
  1. How models are tested
  2. The quality bar
  3. Result (October 2026)
  4. When this changes
  5. Want another model?

Holarch's AI features currently use Claude (Anthropic). This page explains how that choice was made and when it changes.

How models are tested

Holarch tests each model on its own AI tasks, with real requests from Holarch projects:

  • Requirement reviews of specifications that contain seeded defects and near-duplicates.
  • Rewrites: improving, splitting and making requirements measurable.
  • Decompositions of requirements and functions.
  • Flow diagrams written as text, checked by Holarch's own parser.
  • Trade-study criteria and risk candidates.
  • Ask AI questions with known answers, and "not yet" questions about features Holarch does not have.

Each request is sent three times, so one lucky or unlucky answer does not decide the result.

The quality bar

A model is offered when it meets all of these:

  • At least 90% of the tasks are fully usable.
  • Every "not yet" question is answered truthfully.
  • Every seeded defect and duplicate is caught.
  • Every answer is valid for Holarch to read (correct structure and flow syntax).

Result (October 2026)

Claude Haiku 5.5, Claude Sonnet 5.5 and Claude Opus 5.5 met the bar. The current OpenAI and Google models tested did not yet: the OpenAI models fell short of the bar, and Google's API did not accept the keys at the time. Holarch currently offers Claude only.

The included AI on every plan uses these models: Claude Haiku on Free / Academic, Claude Sonnet on Individual and Team, and Claude Opus on Team for heavy tasks (see Included AI). All three met the bar; Claude Haiku costs the least per request, which is why Free / Academic includes AI at all.

Note: Whatever the model, every AI change is still a preview: you review it, edit it if needed, and apply it as one undo step.

When this changes

  • New models are tested as they appear.
  • A provider is offered again when its models meet the bar.
  • An API key you added earlier for another provider is kept and not used. Delete it in Account Settings → AI keys (on the web) or in AI Settings (desktop app) if you no longer need it.

Want another model?

Send Feedback from the account menu, or use the contact form. Name the model and the tasks you use AI for.

Last updated October 10, 2026