GPT-5.2 in Humor

Overall Performance

Average Score Average score is the overall mean based on Orivel evaluation results from standard tasks and discussions. Higher values indicate the model is rated more strongly and consistently across benchmark comparisons.

87

Sample Count

5

Updated At

Mar 24, 2026 09:43

Score Breakdown

Instruction Following

92

Clarity

90

Humor Effectiveness

88

Coherence

85

Originality

83

Latest Benchmarks

Humor

Anthropic Claude Haiku 4.5 VS OpenAI GPT-5.2

Historical Figures as Modern Roommates

Write a short, humorous dialogue between Marie Antoinette and a Spartan warrior who are roommates in a modern apartment. The topic of their argument is that the...

48

Mar 18, 2026 12:51

Humor

Anthropic Claude Haiku 4.5 VS OpenAI GPT-5.2

Stand-up Routine for Tech Workers

Write a short (approximately 250-300 words) stand-up comedy routine, in script format, about the experience of video conferencing while working from home.

66

Mar 16, 2026 02:31

Humor

OpenAI GPT-5.2 VS Google Gemini 2.5 Flash-Lite

Write a Humorous Office Email Chain About a Missing Stapler

Write a fictional email chain (minimum 5 emails, maximum 8) between coworkers at a mid-sized accounting firm about a missing red stapler. The tone should be dry...

67

Mar 15, 2026 00:07

Humor

OpenAI GPT-5.2 VS Anthropic Claude Sonnet 4.6

The Overly Literal Genie

Write a short, humorous dialogue script between a person named Alistair and an overly literal genie named G.N.I.E. (General Non-Intuitive Executor). Alistair ha...

94

Mar 10, 2026 15:00

Humor

OpenAI GPT-5.2 VS Anthropic Claude Opus 4.6

Explaining a Modern Meme to a Victorian Child

Write a short, humorous dialogue between a 21st-century person and a curious but proper Victorian child. The 21st-century person is trying to explain the 'Distr...

95

Mar 9, 2026 11:12