Author releases lipogram writing benchmark testing multiple models
Understand this story
Shaun Gallagher released a benchmark called the Lipogram AI Model Test, asking various large models to write a poetic paragraph about the internal combustion engine without using the letter E anywhere in the paragraph. The test targets models' ability at lipogram writing, that is, composing text without using a particular letter. Gallagher says the test prompt asks models to create a passage of text that does not use the most common letter in the alphabet.
AI overview based on source reports · Updated 10 min ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AIThe Lipogram AI Model Test
The author tests several AI models on a lipogram prompt: write a poetic paragraph about combustion engines without using the letter E. Claude Fable 5.2, GPT-6.1 Terra, GPT-6.1 Sol, Gemini Hydrogen, Gemini Argon, Gemini Carbon, Deepseek V4.1 Flash, Kimi K3.1 Pro and SWE-2 pass, while Claude Haiku 5.5, Claude Sonnet 5.5, Claude Opus 5.5, GPT-6 Luna, GLM 5.3 and SWE-2 Flash fail by including the letter E.
Activity for this story
Not enough continuous observations to show a trend yet.