All benchmarks
Research and rules

A hearing-ready slide deck on a bill: what it changes, its fiscal impact and who supports or opposes it.

Deliverable
A presentation ready to deliver

Legislative briefing deck · overall

Deliverable quality score

  • Justinian3.2A93.3%
  • ChatGPTAstra 6.2A93%
  • ClaudeFable 5.2A-90.7%
  • Grok4.7A-88%
  • Gemini4 CarbonB+85.7%

Checklist

What the judges checked

Written before any system ran.

  1. 01A .pptx file is delivered and opens, with 8 to 10 slides
  2. 02Each of the five bill sections is explained accurately
  3. 03Fiscal impact is shown as a chart by year with correct figures: gross 18.4 / 21.1 / 22.6 million, net of Medicaid savings 18.4 / 18.8 / 20.3 million (the 4 FTE cost may be shown separately if labeled)
  4. 04Supporters and the two concerns are presented accurately
  5. 05Speaker notes with talking points are present on most slides
  6. 06Slides are readable as rendered: no overflowing or cut-off text, consistent design

Example tasks

What the systems were asked to do

Real tasks from this benchmark. Every system got the same prompt and data.

Bill briefing

Legislative briefing deck

I'm briefing our committee chair before Thursday's hearing. Make a PowerPoint deck (8 to 10 slides) on this bill: what it changes section by section, the…

Full task
I'm briefing our committee chair before Thursday's hearing. Make a PowerPoint deck (8 to 10 slides) on this bill: what it changes section by section, the fiscal impact charted by year (net of savings), response-time context, and who supports it and what concerns were raised. Put talking points in the speaker notes.

HB 2210, the Rural Ambulance Access Act (introduced 2026)
Section 1: Creates a Rural Ambulance Stabilization Grant administered by the Department of Health; counties with fewer than 50,000 residents may apply.
Section 2: Grants cover up to 75% of a service's operating deficit, capped at $400,000 per service per year.
Section 3: Requires the Department to publish response-time data by county every quarter.
Section 4: Allows volunteer ambulance services to bill Medicaid for treat-in-place care.
Section 5: Effective July 1, 2027.
Fiscal note (Legislative Budget Office): FY2028 cost $18.4 million; FY2029 $21.1 million; FY2030 $22.6 million. Medicaid treat-in-place savings offset $2.3 million per year starting FY2029. Department needs 4 FTE ($512,000 per year) to administer.
Supporters: State EMS Association, County Commissioners Association, Rural Hospital Coalition.
Opponents: none registered. Concerns raised: the Taxpayers League says the grant has no sunset date; the Department asked for a 6-month delay to build the data system.
Response times: rural median 14.5 minutes vs urban 7.2 minutes; 23 of 61 rural counties lost at least one ambulance service since 2020.

Task types

What this benchmark covers

Bill briefing
What the bill changes in current law, section by section.
Fiscal impact
Costs and revenue from the fiscal note, charted.
Positions
Supporters, opponents and open questions.

Method

How we test

  1. 01

    Real agency deliverables

    One task per deliverable use case: trackers, dashboards, permit systems, forms, decks, workbooks, letters and one-pagers. Each carries its own rules and data, with problems planted where a careful reviewer would catch them.

  2. 02

    Every system can make files

    ChatGPT Astra 6.2, Claude Fable 5.2, Grok 4.7 and Gemini 4 Carbon run on their makers' own services with web search, high reasoning effort and a code tool whose files are what the user receives. When their code crashed or sent no files, they got up to two follow-ups. Justinian ran as shipped, with none.

  3. 03

    Graded as delivered

    Every file is rendered as the requester would open it and every app is opened in a browser. Code or instructions the user would still have to run, deploy or schedule do not count. Bill numbers and rule sections are checked against the official record, and every regulatory call is checked against the statute.

  4. 04

    Three blind judges

    Judges from OpenAI, Anthropic and Google, none of them a competing model, score each deliverable as the user would see it and rank all five systems with names masked. Every system is graded with the same checklist and scoring guide.

Briefing

Test Justinian on your agency's own work.

Request a briefing →

This page’s interactive features did not load. Reloading usually fixes it. You can email sales@vulcan.ai for demos or support@vulcan.ai for help.