GPT-5.5 Instant System Card

OpenAI has released the system card for GPT-5.5 Instant (gpt-5.5-instant), a new model in the Instant series designed for high-speed performance. This release is significant because it is the first model in the Instant series to be classified as having "High capability" in the Cybersecurity and Biological & Chemical Preparedness categories, necessitating the implementation of specific safeguards.

Safety Classifications and Safeguards

GPT-5.5 Instant is treated as a High capability model in two critical safety domains: Cybersecurity and Biological & Chemical Preparedness. While the overall safety mitigation approach for GPT-5.5 Instant remains consistent with previous models in the Instant series, the High capability classification triggers the application of appropriate safeguards to mitigate risks associated with these specific domains.

Model Versioning and Baselines

To provide clarity on the model's positioning within the OpenAI ecosystem, OpenAI has specified the following versioning details:

• Baseline Model: The primary model for baselining against is GPT-5.3 Instant.
• Version Gap: There is no model designated as GPT-5.4 Instant.
• Naming Convention: To avoid confusion between the two 5.5 versions, OpenAI refers to GPT-5.5 as "GPT-5.5 Thinking" to distinguish it from the "Instant" model.

Related Research and Advancements

Beyond the GPT-5.5 Instant release, OpenAI's broader research trajectory includes advancements in agentic AI for scientific computing and safety and alignment for long-horizon models, as indicated by their recent publications.