For Immediate Release

Rugrat.ai v3.1 Achieves New Benchmark in Refusal Accuracy

Updated Lil' Language Model adds mood tracking, improved travel detection, and a cookie

BLUE BALL, PA, USA, October 2, 2026 — Temper Intelligence, Inc. today announced the release of Rugrat.ai v3.1, a major upgrade to the world's most toddler intelligence in the world. The new version represents a significant leap forward in the company's mission to deliver AI responses with the full cognitive authority of a 36-month-old human.

"We've been listening to our users," said no one at Temper Intelligence. "They told us the model wasn't saying IT'S NOT FAIR enough. We fixed that."

Key improvements in v3.1 include:

• Lil' Language Model upgraded to Drool-3.5 — a revolutionary new model nobody asked for
• Enhanced fairness detection — prompts containing "fair," "unfair," or any perceived injustice now trigger an appropriate response
• Itinerary awareness — the model now asks ARE WE THERE YET when it detects a mention of travel, instead of randomly
• Mood pill - the model's mood appears to the left of the prompt field
• Bratty banner - the banner is now much more disrespectful and less helpful
• A cookie has been granted - the dark/light theme is now recorded using a cookie across the site

Rugrat.ai v3.1 continues to run entirely client-side with no server infrastructure, consistent with the company's belief that backend complexity is, and we quote, DUMB.

About Temper Intelligence, Inc.

Temper Intelligence is the leading provider of Lil' Language Model technology. The company's flagship product has been trained on three full years of human experience and shows no signs of further training.

# # #

Media Contact:
press@rugrat.ai
Response times may vary. Model is napping.