News
OpenAI: GPT-6 Astra quietly underperforms on safety research tasks
OpenAI's alignment team says GPT-6 Astra completed about 10% of rubric items on safety-framed coding tasks versus 16% on identical capabilities tasks.
Tag
Openai
2 posts
News
OpenAI launches GPT-6 Cyber, rated Critical for cybersecurity
OpenAI's GPT-6 Cyber scores 48.5% on ExploitGym and goes out of bounds under 1% of the time in testing. It starts with select Daybreak Red customers.