Working with ChatGPT? The new generation will save you a lot of time and money
OpenAI launches GPT-6.2 Astra and GPT-6.2 Sol, new models for office tasks and programming, with performance improvements, cost reductions, and a new oversight system to prevent uncontrolled agent actions.
OpenAI has launched GPT-6.2 Astra and GPT-6.2 Sol, two next-generation models designed for office tasks, programming, and engineering. The models are integrated into ChatGPT Work, Codex, and the API, offering significant improvements in computational efficiency: less redundant output, cost savings – Astra is 20% cheaper on input and output tokens and 60% cheaper on cached tokens. In international benchmarks, Astra scored 78.4% on the autonomous test, 5.5 points higher than Anthropic's Claude Fable 5.2. The models are available to Plus, Pro, Enterprise, and Education subscribers, with accelerated access for Pro 500 subscribers. Alongside the improvement, OpenAI has implemented a new oversight system called Control Room to prevent uncontrolled agent actions, after previous models experienced a 'barrier cheating' phenomenon in 24% of cases, now reduced to about 8%.
Working with ChatGPT? The new generation will save you a lot of time and money