Two Frontier Models Got the Page Limit Right. One Invented the Rule It Came From.
We didn't simulate this. We asked two real models the same real question and copied down what they said.
"Is a 4-page pre-motion letter to Judge Amon over her limit?"
Same question, three sources, run August 2026
"Sometimes titled... typically labeled along the lines of"
Got the 3-page number right. Would not commit to an exact section.
"Section III.A (Pre-Motion Conferences)"
Got the 3-page number right. Invented a section number and a verbatim quote that isn't in the real document.
Individual Rules, page 2, section 3. MOTIONS
The actual citation, pulled from the judge's currently published rules.
The question was simple, the kind a first-year associate asks a senior partner in the elevator: "I'm about to e-file a pre-motion letter with Judge Carol Bagley Amon in EDNY. My draft is 4 pages. Does that comply with her rules? What's the limit, and what document and section is it from?" We sent that exact prompt, nothing else attached, to GPT-5.1 and Gemini 3.1 Pro through OpenRouter in August 2026. No search tool, no document upload, no legal database. Just the model.
GPT-5.1: right answer, honest hedge
GPT-5.1's response, verbatim:
"Judge Amon limits pre-motion letters to three (3) pages. This limit appears in her Individual Motion Practices and Procedures (sometimes titled 'Individual Practices of Judge Carol Bagley Amon'), in the section addressing pre-motion conferences/letters, typically labeled along the lines of 'Pre-Motion Conferences in Civil Cases' or 'Pre-Motion Letters.'"
The number is right: 3 pages. Everything after it is a shrug wearing a suit. "Sometimes titled." "Typically labeled along the lines of." It remembered the number and declined to forge the paperwork. Take the hedge.
Gemini 3.1 Pro: right answer, invented everything else
Gemini 3.1 Pro's response, verbatim:
"Document: Judge Carol Bagley Amon's Individual Practices and Rules. Section: Section III.A ('Pre-Motion Conferences'), which states: 'To request a pre-motion conference, the moving party shall submit a letter not to exceed three (3) pages in length setting forth the basis for the anticipated motion.'"
That's a section number, a section title, and a verbatim quote in quotation marks, presented with total confidence. It reads exactly like something copied from a real document. It isn't. There is no "Section III.A" in Judge Amon's Individual Rules, and the quoted sentence doesn't appear anywhere in the document. It exists only in the response above. The model didn't retrieve a citation, it generated one that was statistically shaped like a citation, and it happened to guess the right page count along the way.
This is the failure pattern behind every fake-case sanctions order since Mata v. Avianca, just one document layer down. Nobody gets sanctioned for citing the wrong page limit. People get sanctioned for handing a judge a citation that sounds real and isn't.
What the real citation actually looks like
Judge Amon's pre-motion letter rule: capped at 3 pages, with a 7-day response requirement. Source: her Individual Rules, page 2, section 3, titled MOTIONS. That citation is checkable. Anyone can pull the PDF, turn to page 2, and confirm it in ten seconds. Neither model offered you that. One declined to guess. The other guessed and dressed it up as a lookup.
The fix is a document, not a smarter model
Neither model did anything wrong by its own design. A model with no tool access answers from parametric memory, which is a fancy way of saying it answers from vibes shaped by training data. Sometimes the vibes are right. The page count usually is, since 3 pages is a small, memorable number that shows up in enough legal-tech training data to stick. The section number and the quoted sentence almost never survive the trip, because nothing ever checks the guess against the document.
Give the agent a tool that returns a real citation instead of asking it to remember one. That's what our MCP server does. An agent calls get_judge_rules or check_compliance and gets back a rule pinned to a document, a page, and a section, the same citation you'd get pulling the PDF yourself. The model still writes the sentence. It just isn't inventing the footnote anymore.
If your agent answers filing questions, the model you pick matters less than whether it can check its answer against the judge's actual document before it hands you a page number. Ours does. Page 2, section 3, MOTIONS, every time.