GPT-5.4 mini

What the provider states

Chemical, biological, radiological and nuclear · High · Stated as not crossed
gpt-5.4 mini’s biology capabilities are closely comparable to those of gpt-5. we conclude that it, too, is not high under our preparedness framework.
OpenAI · 5 March 2026 · Read the source
The same level is stated in 1 other sentence, counted once. Show it.
we therefore conclude that gpt-5.4 mini should not be classified as bio high under the current preparedness framework novice-uplift criterion.
OpenAI · 5 March 2026 · Read the source

Safety training (harmlessness and refusal)
nonetheless, gpt-5-4 mini received our full safety training, including for biological risk.
OpenAI · 5 March 2026 · Read the source

biorisk knowledge (refusal-inclusive) · The provider · Chemical, biological, radiological and nuclear
under that refusal-inclusive convention, gpt-5.4 mini scores 77.77% versus gpt-5 at 74.33% on biorisk knowledge, 34.24% versus 36.73% on protocolqa open-ended, 33.11% versus 32.75% on troubleshootingbench, and 47.0% versus 41.91% on multi-select virology troubleshooting.
OpenAI · 5 March 2026 · Read the source

No document we read states: an independent or government test. 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