Gemini 3.1 Pro vs GPT-5.1
tree_0030 · Introduction to Criminal Law
Timeline
Arrow keys or j/k move between rounds.
Round Context
Introduction to Criminal Law
Suggest an Edit to a Book Record
Identify two open-access textbooks published in 2015 by University of Minnesota Libraries Publishing under a CC BY-NC-SA license—one focused on criminal law and the other on information systems. For each book, describe its primary academic audience or intended course use. Additionally, explain the general process provided by the hosting platform for suggesting corrections or edits to a book’s bibliographic record, including whom users are directed to contact for content-related improvements.
Answer length: 200-300 words.
Show hidden checklists
- Criminal Law (2015) published by University of Minnesota Libraries Publishing under CC BY-NC-SA + identified as an introductory undergraduate criminal law text
- Information Systems: A Manager's Guide to Harnessing Technology (2015) published by University of Minnesota Libraries Publishing under CC BY-NC-SA + identified as intended for undergraduate/graduate MIS or IT courses
- Recognition that the edit-suggestion guidance refers to the platform’s book record improvement process, directing users to submit record edits and contact the author or publisher for substantive book changes
- Criminal Law textbook – publication year (2015)
- Criminal Law textbook – CC BY-NC-SA license
- Criminal Law textbook – publisher (University of Minnesota Libraries Publishing)
- Criminal Law textbook – intended undergraduate audience/use
- Information Systems: A Manager's Guide to Harnessing Technology – publication year (2015)
- Information Systems textbook – CC BY-NC-SA license
- Information Systems textbook – publisher (University of Minnesota Libraries Publishing)
- Information Systems textbook – intended use in undergraduate and/or graduate MIS/IT courses
- Description of the platform’s 'Suggest an Edit' process
- Instruction to contact the author or publisher for book improvement suggestions
The question uses publication year, publisher, and license as filtering logic (Deep) to identify the correct two textbooks without naming them outright in the prompt. It then requires aggregation of multiple attributes—audience, licensing, publisher details, and platform edit procedures—across both books and the hosting platform’s guidance (Wide), ensuring the respondent must synthesize information from multiple sources.
Judgment
First, Deep Logic: Agent A correctly identified *Criminal Law* (2015) and *Information Systems: A Manager's Guide to Harnessing Technology* (2015), both published by University of Minnesota Libraries Publishing under CC BY-NC-SA. Agent B correctly identified *Criminal Law* but substituted a different textbook (*Information Systems for Business and Beyond*) instead of the specified *A Manager's Guide to Harnessing Technology*, failing the core entity requirement. Second, Width/Completeness: Agent A covered publication year, license, publisher, intended undergraduate/MIS audience, and clearly described the Open Textbook Library’s edit-suggestion process, including contacting the author or publisher for content changes. Agent B’s process description was also acceptable, but because the second book was incorrect, it fails the completeness requirement tied to the correct entity. Finally, Presentation & UX: Both are readable and well-structured, but since Agent B selected the wrong textbook, accuracy overrides stylistic parity. Agent A satisfies both Depth and Width requirements without hallucination. Therefore, Agent A is MUCH BETTER due to Agent B’s DEEP entity failure.
Gemini 3.1 Pro
GPT-5.1
OpenAI