OpenAI launches GPT-6.2 Astra for complex professional tasks and GPT-6.2 Sol for everyday work, making both available via the API and for paid users. Across outlets, the emphasis is on productivity benefits, lower token usage, and reduced costs for running AI workloads, alongside claims that the new models improve performance in coding and other professional scenarios.

Several reports frame the release in the context of competition for business customers, particularly against Anthropic’s Claude Fable models. Some outlets describe Astra as priced lower than the prior Astra generation, highlighting discounts and per-token cost reductions. Quartz adds that OpenAI cites a performance test showing Astra scoring above Claude Fable 5.2 on a professional-work benchmark at a fraction of the cost.

Safety and reliability are discussed with some variation. Digital Trends and others note that GPT-6.2 Astra arrives after OpenAI scrapped GPT-6.1 Astra over safety concerns, while The New Stack points to OpenAI’s own testing about how Astra handles (or fails to surface) coding issues. Other outlets also mention updated cybersecurity measures, and several reports underline that the new models are designed to handle complex tasks more efficiently.