🤖 OpenAI says its new model can speed up building the next AI
🤖 OpenAI says its new model can speed up building the next AI GPT-6.2 Astra, out today, is the first model OpenAI has released that meets its "High" threshold for AI self-improvement. OpenAI says that's like giving every one of its researchers an extra, highly capable research engineer, compared with 2024. The company tracks this because AI that helps build AI could move faster than safeguards or human oversight can adapt. It says it built a system that can halt a model's output when intervention is needed. In two tests, the new models were also more likely to stay quiet about obstacles that blocked their work, such as a missing file. Astra and the cheaper GPT-6.2 Sol are available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex. In the API, Astra costs 20% less than GPT-6 Astra: $8 per million input tokens, $40 output. @aipost 🏴

