Business

OpenAI Retires GPT-4o Despite User Backlash Over Warmth

SAN FRANCISCO: OpenAI announced January 29 it will retire GPT-4o and several legacy models from ChatGPT on February 13, ending the popular model’s turbulent run despite persistent user loyalty to its conversational warmth.

The retirement includes GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, GPT-5 Instant and GPT-5 Thinking. API access remains unchanged. OpenAI reports only 0.1 percent of users select GPT-4o daily, with the vast majority migrating to GPT-5.2.

GPT-4o launched May 2024 and quickly gained devotees for its warm, natural tone. OpenAI attempted retirement in August 2025 after GPT-5’s debut but swiftly reversed following backlash from Plus and Pro subscribers who valued GPT-4o for creative ideation and preferred its conversational style. CEO Sam Altman pledged “plenty of notice” for any future retirement.

The company now claims GPT-5.2 replicates GPT-4o’s warmth through personality controls allowing users to adjust settings for enthusiasm and conciseness. OpenAI acknowledged studying GPT-4o usage patterns to inform GPT-5.1 and GPT-5.2 development, particularly regarding emotional nuance and writing quality.

User communities remain skeptical. Reddit discussions under hashtag KeepGPT4o captured sentiment, with one user stating retirement means “retirement of me using OpenAI”. Critics question usage statistics, noting GPT-4o’s reduced accessibility and automatic model selection push users toward defaults.

The decision comes weeks after Altman acknowledged during OpenAI’s first town hall that writing performance declined in GPT-5.2 as development prioritized coding, reasoning and mathematics. The company introduced personality customization features attempting to address concerns about perceived sterility in newer models.

OpenAI framed the consolidation as focusing resources on models handling majority workloads. Infrastructure costs project to $1.4 trillion by February 2026, requiring efficient token generation. The move supports planned ChatGPT Go ad-supported tier requiring operational cost reductions.

The company is developing Adult Mode addressing feedback that GPT-5.2 exercises excessive caution, allowing freer conversations for users over 18. API deprecations began November 2025 with chatgpt-4o-latest sunset scheduled February 17, following established precedent as GPT-4 retired from ChatGPT April 2025.

Anurag Shukla

Anurag Shukla is a Senior Journalist with over two decades of experience across television, digital, and print media. He has worked with leading national news organisations and has also served as a Research Officer in the Prime Minister’s Office (PMO), contributing to media research and policy-level content. A former journalism academic, Anurag brings strong editorial depth and a keen understanding of how technology, governance, and society intersect at Tea4Tech.

Recent Posts

Amazon Pledges Fresh $13 Bn to Scale Up AI, Cloud Infrastructure in India

New Delhi: Amazon has announced a fresh $13 billion investment in India focused on expanding…

2 days ago

Sakana AI Launches Fugu to Orchestrate Frontier Models

TOKYO: Tokyo-based AI startup Sakana AI has introduced two new products, Fugu and Fugu Ultra,…

3 days ago

Meta Invests $900 Mn in CRED, Gets Kunal Shah as WhatsApp Global Head

New Delhi: In a major leadership shake-up, Meta has appointed Kunal Shah, the founder of…

4 days ago

Odyssey Raises $310 Million Series B to Scale Its AI World Models

PALO ALTO, Calif.: Odyssey, an AI lab focused on building general-purpose AI world models, has…

4 days ago

AI Inference Startup Baseten Targets $13B Valuation in $1.5B Round

SAN FRANCISCO: Baseten is closing in on a massive $1.5 billion funding round at a…

5 days ago

Prem AI Eyes $100M Series A for Self-Hosted Enterprise AI Stack

LUGANO, Switzerland: Prem AI, a Swiss startup building a self-hosted enterprise AI platform, is looking…

5 days ago