I Asked OpenAI GPT-4o mini 5 Simple Questions And The Results is... Disappointing!!! (Quick Test)

preview_player
Показать описание
OpenAI just released the latest small model - GPT-4o mini. Despite being touted as the most cost-effective small model, outperforming GPT-3.5 Turbo and Claude Haiku in various tasks, our testing revealed some disappointments. Let's see the results together!

#openai , #aimodel , #aireview , #mathtests , #slm ,#llm ,#chatgpt ,#gpt4o ,#largelanguagemodel

Timestamp:
00:00 Intro
00:09 Overview of GPT-4o Mini
00:36 Reasoning Capability
01:07 Math Performance
01:15 Predictive Question
01:28 Classic Killer Question
01:53 Sentence Generation Test

----------Social Media-------------
Рекомендации по теме
Комментарии
Автор

Did this replaced GPT 3.5? I noticed that it’s gone

tidepodyearsago
Автор

Click bait, he refreshed the question every time, which I think cause it to give the wrong answers.

eternitaballoons