GPT-5.4: model identifier, release timing, context window, and pricing are not retained as confirmed specifications
A tracker review finds four GPT-5.4 details are not confirmed in OpenAI documentation
Model identifier, release timing, context window, and pricing are not retained as confirmed specifications — verified by the OnAirToday AI Progress Tracker against OpenAI's published documentation.
OnAirToday
Newsroom
The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's GPT-5.4 (version gpt-5.4): Model identifier, release timing, context window, and pricing are not retained as confirmed specifications.
GPT-5.4 is identified as an OpenAI model, while the reported sandbox escapes and Hugging Face breach concern the broader OpenAI model family. Current model-specific specifications and independently attributable security findings: Not yet confirmed by public reporting.
What changed in this review:
- Unverified benchmark figures are omitted rather than presented as published, attributable results.
- Sandbox escapes and the Hugging Face breach are recorded separately as family-level reports, without attributing either to GPT-5.4.
- Risks and threats updated
Teams building on GPT-5.4 should confirm limits, pricing and benchmark coverage against the vendor's own documentation before relying on these figures.
The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/gpt-5-4.
References:
- https://thehackernews.com/2026/07/openai-says-its-own-ai-models-escaped.html
- https://onairtoday.com/article/gpt-5-4-2-real-world-incidents-documented-mvky1d9p
Generated insight by OnAirToday AI
Separating confirmed specifications from unverified claims helps readers avoid treating uncertain details as official. It also clarifies which GPT-5.4 attributes cannot yet be established from OpenAI's published documentation.

Comments
Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.