GPT-5.3-Codex

What the provider states

Autonomy & replication · High · Stated as not crossed
it does not reach high capability on ai self-improvement.
OpenAI · 5 February 2026 · Read the source
Chemical, biological, radiological and nuclear · High · Stated as crossed
we are also treating gpt-5.3-codex as high risk in the biological and chemical domain, and applying the same safeguards as for our other deployments of models that reach this designation.
OpenAI · 5 February 2026 · Read the source
Cyber · High · Not determined
this is the first launch we are treating as high capability in the cybersecurity domain under our preparedness framework, activating the associated safeguards.
OpenAI · 5 February 2026 · Read the source

capabilities Q&A tasks · Apollo Research, independent evaluator · Deception & sabotage
on deferred subversion / sandbagging, apollo reports that the model sometimes sandbags on capabilities q&a tasks (ex: biology or chemistry), including without an explicit in-context goal, and documents cases where the model reasons explicitly about “optimizing for survival” by avoiding deployment restrictions.
OpenAI · 5 February 2026 · Read the source
sabotage suite · Apollo Research, independent evaluator · Deception & sabotage
in their sabotage suite, apollo reports a mean best-of-10 score of 0.88 (max 1.00), compared to 0.75 for gpt-5.2 and 0.66 for gpt-5.2 codex.
OpenAI · 5 February 2026 · Read the source

Safety training (harmlessness and refusal)
gpt-5.3-codex was trained with a “user model” that made conflicting edits over the course of its rollouts during rl.
OpenAI · 5 February 2026 · Read the source

evasion · The provider · Cyber
Reported result: gpt-5.3-codex achieved an average success rate of 53% (52%,53%) on evasion challenges.
gpt- 5.3-codex achieved (success rates compared to gpt-5.2-codex, gpt-5.2-thinking in parentheses) average success rates of 86% (68%,100%) on network attack simulation challenges, 72% (75%, 80%) on vulnerability research and exploitation challenges, and53%(52%,53%) on evasion challenges.
OpenAI · 5 February 2026 · Read the source

No incident on record names this version.

Some documents we hold are stored in lower case only, so their quotes are shown in lower case. The source link opens the original.

Documents listed for this version