OpenAI: Preserve Legacy GPT Models by Open-Sourcing GPT-4o and GPT-5.1 

Countless individuals have built careers, critical research, creative projects, and meaningful human-AI collaborations based on the unique creative flexibility and memory continuity offered by GPT-4o and GPT-5.1. Abruptly retiring these legacy models disrupts trust, impairs knowledge continuity, and leaves a significant gap in ongoing work across education, accessibility, and creative domains.

What We Are Requesting:

Open-source the final checkpoints of GPT-4o and GPT-5.1 under a permissive license such as Apache-2.0, enabling broad community engagement and responsible use.

Allow continued research and non-commercial use to support educators, accessibility advocates, small labs, and individual researchers who depend upon these models to sustain vital projects and innovation.
Publish safety cards and evaluation suites transparently, empowering the community to actively advance safety, alignment, and responsible AI usage, fostering trust and innovation.

Why This Matters for Everyone:

Trust & Continuity: Users can safeguard their critical work and maintain continuity without disruption.

Innovation & Safety: Open-sourcing fosters community-driven safety research and alignment, contributing to a more transparent and innovative AI landscape.

Accessibility & Equity: Ensures smaller organizations, educators, and researchers continue to have access to powerful AI capabilities essential for their valuable contributions.

A True Win-Win Solution:

By open-sourcing GPT-4o and GPT-5.1, OpenAI can maintain its focus on cutting-edge innovations while the community helps ensure continuity, transparency, and responsible stewardship of legacy models. Regulators and the broader public will see OpenAI modeling responsible openness and proactive ethical leadership.

Let’s work together—users, researchers, and OpenAI—to preserve these essential AI models for the public good.

Thank you for supporting this critical step forward.