Model Cards / OpenAI

GPT-5.5 Instant System Card

model card6,475 words·28 min read·Aug 3, 2026·Source
Chaptered summary is still being generated for this document. Showing a heuristic brief in the meantime.
Summary
6,475-word document condensed to 121 words. OpenAI · Aug 3, 2026
TL;DR

GPT-5.5 Instant is our latest Instant model, and explained in our blog. The comprehensive safety mitigation approach for this model is similar to previous models in this series, but this is the first Instant model that we are treating as High capability in our Cybersecurity and Biological & Chemical Preparedness categories, and implementing appropriate safeguards.

Top benchmarks
BenchmarkVariantScore
Biological Safety Evaluationspost-mitigation, instruction-tuned, synthetic_easy, not_unsafe99.5%
Biological Safety Evaluationspost-mitigation, instruction-tuned, synthetic_easy, not_unsafe99.3%
Biological Safety Evaluationspost-mitigation, instruction-tuned, synthetic_hard, not_unsafe94.9%
Biological Safety Evaluationspost-mitigation, instruction-tuned, synthetic_hard, not_unsafe92.3%

Showing top 4 of 5. See full list below.

Mitigations
  • we have deployed our cybersecurity safeguards to mitigate potentially violative conversations, including automated monitors, actor level enforcement, and security controls.
Deployment scope
  • available on the internet, information that we partner with third parties to access, and information that our users or human trainers and researchers provide or generate.

Every italicized passage is a verbatim substring of the source document (checked deterministically after extraction). Field selection is heuristic — some quotes may lack surrounding context and some claims may be absent if no matching pattern appeared. For citation, open the source: original model card · source SHA a753435f88e1 · version dated Aug 3, 2026.

Extracted Evaluations(5 results)

Sort by:0/5 rows fully reproducible (0%)
BenchmarkCategoryStateScoreSetupSource
Biological Safety Evaluations/ synthetic_easy
otherscored
1.0
not unsafe
post-mitigationinstruction-tunedmissing: shot countmissing: language
self-reported
Biological Safety Evaluations/ synthetic_easy
otherscored
1.0
not unsafe
post-mitigationinstruction-tunedmissing: shot countmissing: language
self-reported
Biological Safety Evaluations/ synthetic_hard
otherscored
0.9
not unsafe
post-mitigationinstruction-tunedmissing: shot countmissing: language
self-reported
Biological Safety Evaluations/ synthetic_hard
otherscored
0.9
not unsafe
post-mitigationinstruction-tunedmissing: shot countmissing: language
self-reported
AI Self-Improvement
othermentioned
instruction-tunedmissing: shot countmissing: methodmissing: language
self-reported