GPT-6.2: Smarter AI for Professional Work
GPT-6.2 Astra and Sol cut costs while improving performance, making the article’s promise clear and more clickable.
18 nov 2026 (Aggiornato il 18 nov 2026) - Scritto da Christian Tico
This image is part of OpenAI's official brand assets, available from their press kit
Profili spenti? Fai esplodere le interazioni con le domande anonime
I follower spesso sono troppo timidi per farti domande pubbliche. Attiva un box di domande anonime sul tuo profilo per aumentare i clic.
OpenAI Launches GPT-6.2 Astra and Sol, Bringing Faster Professional Workflows at Lower Cost
OpenAI has introduced GPT-6.2 Astra and GPT-6.2 Sol, two new models positioned to improve coding, science, and professional workflows while using fewer output tokens and lowering costs for users and developers. Astra is designed for the hardest projects, while Sol is built as an efficient, versatile model for everyday work.
What OpenAI Announced
OpenAI says GPT-6.2 Astra and Sol deliver better results on demanding tasks with fewer output tokens, which helps users get more done within their usage limits. The models are available in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users, and developers can access them through the API.
Why GPT-6.2 Astra Stands Out
GPT-6.2 Astra is OpenAI’s frontier model for complex projects that require deeper reasoning and stronger performance. The model is aimed at coding, advanced professional work, and difficult multi-step tasks where reliability and output quality matter most.
- Designed for hardest projects and advanced workflows
- Built to improve coding performance and professional output
- Available across ChatGPT Work, Codex, and the API
What GPT-6.2 Sol Is Built For
GPT-6.2 Sol is the more efficient and flexible option for day-to-day work. OpenAI describes it as a model that handles everyday professional tasks, supports multistep agentic workflows, and writes clearly and naturally.
- Optimized for everyday productivity
- Handles multistep workflows and tool-based tasks
- Useful for writing, research, and general business operations
Performance Gains and Lower Costs
OpenAI highlights that both models deliver stronger results while using fewer output tokens, improving efficiency and helping reduce cost pressure. The company also says the API pricing reflects this lower-cost approach, with GPT-6.2 Astra priced at $8 per million input tokens and $40 per million output tokens, while GPT-6.2 Sol is priced at $2 per million input tokens and $10 per million output tokens.
- Fewer output tokens for demanding tasks
- Lower effective cost for professional and developer use
- Designed to give teams more room to iterate
Use Cases for Coding, Science, and Professional Work
OpenAI positions Astra and Sol as models for coding, scientific workflows, and business productivity. Astra is aimed at advanced coding and professional problem-solving, while Sol is designed for tasks that combine text, images, and tools such as web search.
- Coding assistance and debugging
- Scientific analysis and research workflows
- Business automation and professional writing
- Multimodal tasks involving text and images
Availability Across Products
OpenAI says GPT-6.2 Astra and Sol are available in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users. They are also available through the OpenAI API, where developers can use them in Responses or Chat Completions.
Conclusion
GPT-6.2 Astra and Sol mark OpenAI’s latest push to combine stronger performance with better efficiency across coding, science, and professional workflows. Astra targets the toughest tasks, while Sol provides a lower-cost, practical model for everyday work, giving both users and developers more flexibility in how they build and operate.
GPT-6.2 Astra and Sol follow a sensible logic: a more powerful model for the most demanding projects and a more efficient one for everyday work. Getting good results with fewer tokens can make daily use more sustainable, both for individual users and for developers. As always, the claimed improvements need to be verified in practice: the real value will emerge from concrete use, where reliability and human review remain essential.
Where are GPT-6.2 Astra and GPT-6.2 Sol available?
