GPT-5.1 vs Claude Opus 4.1
tree_0024 · 5 Best Practices for Public Relations
Timeline
Arrow keys or j/k move between rounds.
Round Context
5 Best Practices for Public Relations
Apply Now
Identify the public university located in Maryville, Missouri, that publishes a specific guide titled '5 Best Practices for Public Relations' through its online program. Provide a detailed list of these five best practices. Additionally, according to this same institution's resources, what are the specific 'Important Soft Skills' recommended for PR professionals?
Answer length: 200-300 words.
Show hidden checklists
- Target Entity: Northwest Missouri State University (Northwest)
- Logic Proof: Correctly maps 'university in Maryville, Missouri' to the specific '5 Best Practices' guide.
- Best Practice 1 (e.g., Honest communication/Truthfulness)
- Best Practice 2 (e.g., Relationship building)
- Best Practice 3 (e.g., Knowing the audience)
- Best Practice 4 (e.g., Social media utilization)
- Best Practice 5 (e.g., Proactive measures/Crisis management)
- Soft Skill 1 (e.g., Strong writing/communication)
- Soft Skill 2 (e.g., Research/Analysis)
- Soft Skill 3 (e.g., Creativity)
- Soft Skill 4 (e.g., Time management/Organization)
The question utilizes Deep Logic by masking the institution's name with its geographic location (Maryville, Missouri), forcing the agent to reason about the source of the '5 Best Practices' guide. It employs Wide Logic by requiring the aggregation of two distinct sets of information (the 5 practices and the soft skills) derived from the identified source.
Judgment
Both agents correctly identified the university (Northwest Missouri State University). However, the query requested a specific list from a specific guide ('5 Best Practices for Public Relations'). Agent B correctly retrieved 4 out of the 5 specific practices mentioned in the actual university guide (Know Your Audience, Be Proactive, Maintain Media Relationships, Master Social Media). It only missed 'Honesty' (replacing it with Measurement). Agent A, by contrast, hallucinated a completely generic list of PR best practices (Research, Messaging, etc.) that does not match the source text requested. Agent B demonstrated actual retrieval capabilities, while Agent A relied on predictive text generation for a generic topic.
GPT-5.1
OpenAI
Claude Opus 4.1
Anthropic