ChatGPT User? Former OpenAI Researcher Andrej Karpathy Breaks Down Which Version Works Best For Each Task

On Monday, former OpenAI researcher Andrej Karpathy shared a practical breakdown of ChatGPT's various model versions, helping users understand which one to use based on their specific needs.

What Happened: In a post on X, formerly Twitter, Karpathy said many users remain confused about ChatGPT versions like GPT-4o, GPT-4 and "o3."

"o3 is the obvious best thing for important/hard things. It is a reasoning model that is much stronger than 4o and if you are using ChatGPT professionally and not using o3 you’re ngmi," he said.

He explained that GPT-4o, OpenAI's latest model, is ideal as a "daily driver" for easy-to-moderate tasks. "o4 is only available as mini for now, and is not as good as o3, and I’m not super sure why it’s out right now."

See Also: Sundar Pichai Reveals Google-Parent Once Super Intensely Debated About Buying Netflix: ‘In A World Of Butterfly Effects…’

Karpathy offered a quick reference:

• Use 4o for simple questions (~40% of his use)
• Use o3 for complex, high-stakes queries (~40%)
• Use GPT-4.1 for casual coding (~10%)
• Use the "Deep Research" toggle—powered by o3—for deep dives (~10%)

He added that this guidance only applies within ChatGPT, as he also uses Anthropic's Claude, Alphabet Inc.'s (NASDAQ:GOOG) (NASDAQ:GOOGL) Google Gemini, xAI's Grok and Perplexity for other tasks.

Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.

Why It's Important: OpenAI is reportedly expected to generate over $12 billion in total revenue by the end of 2025, up from $3.7 billion in 2024.

While the Sam Altman-led company hasn’t revealed revenue specifically from its enterprise segment, a Reuters report noted that paying seats for its ChatGPT Enterprise offering rose to 3 million in May, up from 2 million in February.

Photo Courtesy: Primakov On Shutterstock.com

Check out more of Benzinga’s Consumer Tech coverage by following this link.

Read Next:

• Mark Zuckerberg Warns Of ‘Serious Disadvantage' As China's Data-Center Blitz Could Let DeepSeek Leapfrog US AI Labs

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.