OpenAI’s GPT-5 was supposed to simplify the ChatGPT experience with an intelligent “model router” that automatically picked the best AI for each query. Instead, the company has brought back — and even expanded — the model picker. Users now have “Auto,” “Fast,” and “Thinking” modes for GPT-5, plus access to older favorites like GPT-4o, GPT-4.1, and o3. The move follows backlash from fans of legacy models and performance issues with GPT-5’s router on launch day. OpenAI leaders acknowledge they still need to refine personalization, as human attachment to specific AI models proves stronger than expected.


Key Points:

  • GPT-5 launched as a unified model with an automatic router to remove the need for a model picker.
  • New GPT-5 modes — Auto, Fast, and Thinking — give users manual control again.
  • Paid users can now access older models like GPT-4o, GPT-4.1, and o3, despite their recent deprecation.
  • Backlash followed GPT-4o’s removal, with many users attached to specific models’ personalities.
  • GPT-5’s router had issues at launch, with users reporting weaker performance than earlier models.
  • OpenAI plans to make GPT-5’s personality “warmer” but “less annoying” than GPT-4o.
  • VP of ChatGPT, Nick Turley, says quick iteration is a strength, even if version one isn’t perfect.
  • Aligning AI behavior to individual user preferences remains a difficult challenge.
  • Human attachment to AI models is becoming a real phenomenon — with examples like a “funeral” for Anthropic’s Claude 3.5 Sonnet.

Key Quotes:

  • Sam Altman: “We are working on an update to GPT-5’s personality which should feel warmer than the current personality but not as annoying (to most users) as GPT-4o… we really just need to get to a world with more per-user customization of model personality.”
  • Nick Turley: “We’re not always going to get everything on try #1 but I am very proud of how quickly the team can iterate.”

Implications:

  • GPT-5’s failure to replace the model picker suggests automated AI routing still can’t meet diverse user needs.
  • OpenAI is acknowledging the emotional connection users form with AI personalities, a factor that could shape future product design.
  • The return of legacy models shows the company’s need to balance innovation with customer loyalty.
  • Personalization — not just speed or accuracy — may become the next battleground for AI platform competition.
  • As human-AI relationships deepen, managing the psychological and social impacts of these tools will be increasingly important.

Source: https://techcrunch.com/2025/08/12/chatgpts-model-picker-is-back-and-its-complicated/

Share This Article

Related Post

Google AI Mode Is Here — And It’s Cha

Google is officially rolling out AI Mode, a new feature...

OpenAI Unveils GPT-4.1 Nano – Its Fastest, Smallest, and Cheapest AI Yet

OpenAI Unveils GPT-4.1 Nano – Its Fastest,

OpenAI has just announced the launch of its new GPT-4.1...

Freepik Challenges AI Norms with Licensed-Only Image Generator

Freepik Challenges AI Norms with Licensed-Onl

Freepik has just entered the AI image generation race w...

Leave a Comment

Prove your humanity: 7   +   8   =  

I'm a paid user

(I’ve purchased Next Level Marketing AI credits)

AISQ | Squirrly created this web Customer App for all of you who own licenses for AISQ’s Next Level Marketing AI, AISQBusiness, Squirrly SEO, Hide My WP Ghost and more.

Read More about Customer App by AISQ | Squirrly, on the Squirrly Company’s official website

I'm a free user

(I haven’t purchased any Next Level Marketing AI credits)