OpenAI, creator of popular chatbot ChatGPT, said that GPT-5.4-Cyber will initially be rolled out on a limited basis to vetted security vendors, organisations and researchers because of its more permissive design.
Mythos, announced on April 7, is being deployed as part of Anthropic’s “Project Glasswing”, a controlled initiative under which select organisations are permitted to use the unreleased Claude Mythos Preview model for de…
The company is also expanding its Trusted Access for Cyber programme to thousands of verified individual defenders and hundreds of teams protecting critical software, it said in a post on its website.
Users approved for the highest tier will gain access to GPT-5.4-Cyber, which has fewer restrictions on sensitive cybersecurity tasks such as vulnerability research and analysis.
Ключевые утверждения в источнике B
OpenAI said its goal is to make advanced defensive tools “as widely available as possible while preventing misuse” through automated verification systems rather than manual gatekeeping decisions.
OpenAI said Codex Security has contributed to fixes for more than 3,000 critical and high-severity vulnerabilities across the ecosystem since its recent broader launch.
OpenAI also noted in its announcement that capture-the-flag benchmark performance across its models improved from 27% on GPT-5 in August 2025 to 76% on GPT-5.1-Codex-Max in November 2025 and said it is planning and eval…
OpenAI is pitching the release as preparation for more capable models expected later this year, saying that it’s “fine-tuning our models specifically to enable defensive cybersecurity use cases, starting today with a va…
Текстовые доказательства
Доказательства из источника A
ключевое утверждение
OpenAI, creator of popular chatbot ChatGPT, said that GPT-5.4-Cyber will initially be rolled out on a limited basis to vetted security vendors, organisations and researchers because of its…
Ключевое утверждение, на котором строится подача материала.
ключевое утверждение
Mythos, announced on April 7, is being deployed as part of Anthropic’s “Project Glasswing”, a controlled initiative under which select organisations are permitted to use the unreleased Clau…
Ключевое утверждение, на котором строится подача материала.
Доказательства из источника B
ключевое утверждение
OpenAI is pitching the release as preparation for more capable models expected later this year, saying that it’s “fine-tuning our models specifically to enable defensive cybersecurity use c…
Ключевое утверждение, на котором строится подача материала.
ключевое утверждение
OpenAI said Codex Security has contributed to fixes for more than 3,000 critical and high-severity vulnerabilities across the ecosystem since its recent broader launch.
Ключевое утверждение, на котором строится подача материала.
оценочная маркировка
The new model has been purpose-built to lower refusal boundaries for legitimate cybersecurity tasks, or in the words of OpenAI, is “cyber-permissive” and adds capabilities not available in…
Оценочная маркировка, которая подталкивает к нормативной интерпретации.
Доказательства паттернов искажения
Для этого блока пока не удалось выделить короткие доказательные фрагменты.
Как формируются сигналы оценок
Сигнал оценки искаженийИтоговый сигнал учитывает фрейминг, эмоционально нагруженные формулировки, выборочный акцент и паттерны односторонней подачи.
Сигнал эмоциональностиЭмоциональность повышается, когда в доказательствах чаще встречаются эмоциональная лексика и оценочные маркировки.
Сигнал односторонностиОдносторонность повышается при доминировании одного фрейма и слабом присутствии альтернативных интерпретаций.
Сигнал доказательностиДоказательность растёт при наличии конкретных утверждений, атрибутированных цитат и проверяемых контекстных деталей.