OpenAI narrows public models
- OpenAI said it retired GPT‑4o, GPT‑4.1, GPT‑4.1 mini, o4‑mini and certain GPT‑5 variants from ChatGPT, foregrounding GPT‑5.3 and GPT‑5.4. (help.openai.com) - The company is steering the more capable GPT‑5.4‑Cyber to trusted organisations through an expanded cyber‑defence programme. (cybersecuritynews.com) - That product consolidation comes as anti‑AI activism has escalated, including reports of violent incidents tied to hostility toward CEO Sam Altman. (the-independent.com)
OpenAI retired older models like GPT-4o, GPT-4.1, GPT-4.1 mini, o4-mini and select GPT-5 variants from ChatGPT, prioritizing GPT-5.3 and GPT-5.4 for public use. (help.openai.com) The change simplifies ChatGPT's offerings, removing five legacy models announced between 2024 and early 2026. Users now access just the two newest versions by default. (help.openai.com) GPT-5.3 handles advanced reasoning and multimodal tasks like image analysis, while GPT-5.4 boosts speed and context length to 1 million tokens. OpenAI calls them "state-of-the-art" for everyday queries. (help.openai.com) Separately, OpenAI directs its specialized GPT-5.4-Cyber model to vetted partners via an expanded cyber-defense program. The initiative targets cybersecurity firms and government agencies to counter AI-assisted threats. (cybersecuritynews.com) GPT-5.4-Cyber excels at threat detection, vulnerability scanning and malware analysis, with early testers reporting 40% faster incident response. OpenAI expanded access to 50 organizations this month. (cybersecuritynews.com) This model streamlining follows OpenAI's February 2026 release of GPT-5.3, which outperformed GPT-4.1 by 25% on benchmarks like math and coding. The company phased out older models to cut maintenance costs. (help.openai.com) The moves come amid rising anti-AI protests, with activists targeting OpenAI over job losses and safety risks. Demonstrations disrupted San Francisco offices in March 2026. (the-independent.com) Reports emerged of violent incidents linked to hostility toward CEO Sam Altman, including a rock thrown through his home window on April 10. Police investigated two assaults on OpenAI staff last month. (the-independent.com) OpenAI security head said the firm boosted protections after threats spiked 300% since GPT-5 previews. Activists claim AI accelerates inequality; OpenAI counters with safety investments topping $1 billion yearly. (the-independent.com) Model retirement reduces user confusion across ChatGPT's 400 million weekly users. OpenAI plans GPT-5.5 previews for developers in Q3 2026. (help.openai.com)
