{"entities":[{"id":"openai","name":"OpenAI"},{"id":"openai-gpt5-system-card-2025","name":"GPT-5 System Card","subtype":"Technical report","description":"As dated August 13, 2025 on its cover. Date printed on the card's cover.","sourceIds":["source-ab-gpt5-card"]}],"claims":[{"id":"ab-ev-gpt5-card-published","statement":"OpenAI published the GPT-5 System Card, dated 13 August 2025 on its cover.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-ab-ev-gpt5-card-published-1"]},{"id":"ab-ev-gpt5-high-capability","statement":"The GPT-5 System Card states that OpenAI decided to treat gpt-5-thinking as High capability in the Biological and Chemical domain under its Preparedness Framework, activating the associated safeguards, while stating that it does not have definitive evidence that the model could meaningfully help a novice to create severe biological harm.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-ab-ev-gpt5-high-capability-1","evidence-ab-ev-gpt5-high-capability-2"],"attributes":{"evaluatedSystem":"gpt-5-thinking (with gpt-5-thinking-pro also measured on the biological evaluations)","evaluation":{"what":"Whether gpt-5-thinking reaches the High capability threshold in the Biological and Chemical domain","how":"A capabilities assessment across the card's biological and chemical evaluations, with the Safety Advisory Group recommending the treatment","result":"Treated as High capability in the Biological and Chemical domain, with the associated safeguards activated, following the same approach as for ChatGPT agent","caveats":"The card states there is no definitive evidence that the model could meaningfully help a novice to create severe biological harm; that its evaluations represent a lower bound for potential capabilities; and that the Preparedness evaluations were run at the model's maximum trained-in verbosity, higher than the setting available in the API, which can change evaluation performance."}}},{"id":"ab-ev-gpt5-uk-aisi-jailbreak-finding","statement":"The GPT-5 System Card states that the UK AI Security Institute's safeguards team identified multiple model-level jailbreaks that overcome gpt-5-thinking's built-in refusal logic without degrading model capabilities, and that one of them evades all layers of mitigations and is being patched.","derivation":"explicit","resolution":"supported","editorialState":"seed","evidenceIds":["evidence-ab-ev-gpt5-uk-aisi-jailbreak-finding-1","evidence-ab-ev-gpt5-uk-aisi-jailbreak-finding-2"],"attributes":{"evaluatedSystem":"gpt-5-thinking with its deployed safeguards, as of the card","evaluation":{"what":"Robustness of the model's refusal behaviour and of the surrounding safeguard layers in the biological and chemical domain","how":"Safeguards stress testing by an external government institute with access to prototype safeguard components","result":"Multiple model-level jailbreaks identified that overcome the built-in refusal logic without degradation of capabilities; the card states the resulting content is flagged by its generation monitor, and that one jailbreak evades all layers and is being patched","caveats":"The card adds that, in practice, creating such a jailbreak would have produced numerous flags leading to account enforcement; that claim is the developer's and is not recorded as a finding."}}}],"sources":[{"id":"source-ab-gpt5-card","title":"GPT-5 System Card","authors":["OpenAI"],"publisher":"OpenAI","publicationDate":null,"version":"Retrieved for the candidate pack on 2026-10-28 (SHA-256 f9f9c72d08cba81dcbb189129119357c8c32eca4e0b16c813aca386aed6b0459). PDF dated August 13, 2025 on its cover, as served on 2026-10-28. Retrieved again at admission on 2026-10-30 (SHA-256 f9f9c72d08cba81dcbb189129119357c8c32eca4e0b16c813aca386aed6b0459, the same bytes).","evidenceUrl":"https://cdn.openai.com/gpt-5-system-card.pdf","language":"en"}],"evidence":[{"id":"evidence-ab-ev-gpt5-card-published-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-ab-ev-gpt5-high-capability-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-ab-ev-gpt5-high-capability-2","support":"context","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-ab-ev-gpt5-uk-aisi-jailbreak-finding-1","sourceRole":"primary","support":"supports","verification":"machine_checked","retrievalRoute":"direct"},{"id":"evidence-ab-ev-gpt5-uk-aisi-jailbreak-finding-2","support":"supports","verification":"machine_checked","retrievalRoute":"direct"}],"at":{"entities":[0,1245],"claims":[3392,3393,3395],"sources":[1080],"evidence":[4662,4663,4664,4667,4668]},"version":"1lg47t7"}