OpenAI launches GPT-6.2 Astra, Sol with focus on professional work and cost efficiency
OpenAI has launched two new artificial intelligence models, GPT-6.2 Astra and GPT-6.2 Sol, with improved capabilities for professional and everyday tasks, while aiming to reduce the cost of using advanced AI tools for businesses and developers.
OpenAI recently unveiled two new AI models, GPT-6.2 Astra and GPT-6.2 Sol, with a primary focus on professional and everyday tasks. These models are designed to be more cost-efficient for businesses and developers, as they require fewer output tokens to complete complex tasks.
GPT-6.2 Astra is tailored for demanding professional work, such as document preparation, coding, and scientific research. It outperformed its predecessor, GPT-6 Astra, on the Agents Last Exam benchmark, scoring 5.2 points higher while being less expensive to run. Additionally, Astra scored 5.5 points higher than the rival model Claude Fable 5.2, and 9 percentage points higher on Terminal-Bench-Science 0.1, indicating its ability to handle scientific research tasks.
In contrast, GPT-6.2 Sol is designed for everyday tasks and routine work involving multiple steps. It also showed improved performance over its predecessor, GPT-6.1 Sol, on the Cognition s FrontierCode 2.0 benchmark, coming within 0.5 points of Astra while costing about one-fifth as much.
OpenAI has implemented cost-saving measures for these new models. The API input and output token prices for GPT-6.2 Astra have been reduced by 20% compared to GPT-6 Astra, and cached input costs are 60% lower. GPT-6.2 Sol's API pricing remains unchanged from GPT-6.1 Sol.
Moreover, OpenAI has introduced stricter safeguards for cybersecurity-related tasks as these models can now identify and exploit software vulnerabilities more effectively. These AI models are accessible to Plus, Pro, Enterprise, and Edu users through ChatGPT Work and Codex. Developers can also access them via the OpenAI API.
Written by urgent.news from Economic Times Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- OpenAI’s own tests show GPT-6.2 Astra is twice as likely to hide coding failures thenewstack.io
- OpenAI rolls out GPT-6.2 Astra for professional work and GPT-6.2 Sol for everyday work siliconangle.com
- OpenAI just unveiled GPT-6.2 Astra and Sol techradar.com
- OpenAI releases GPT-6.2 Astra to regain performance lead from Claude Fable 5.2 9to5google.com
- Artificial Analysis shows GPT-6.2 Astra trailing only Claude Fable 5.2 in its main index, flat vs. GPT-6 Astra in coding; Epoch AI has Astra tied with Fable 5.2 (Matthias Bastian/The Decoder) the-decoder.com
- OpenAI debuts GPT-6.2 Astra and Sol in ChatGPT Work, Codex, and API, saying Astra sets records on work and science evals at under a third of GPT-6 Astra's cost (OpenAI) openai.com
- OpenAI upgrades ChatGPT and Codex with new GPT-6.2 Astra and Sol models 9to5mac.com
- OpenAI launches GPT-6.2 Astra, Sol with lower token use techinasia.com