OpenAI unveils GPT-6.2 models for advanced AI tasks

OpenAI has introduced two new artificial intelligence models, GPT-6.2 Astra and GPT-6.2 Sol, as it looks to improve how AI handles professional work, software development and complex digital tasks.

Announced on November 17, the models are designed to deliver better results while using fewer output tokens, helping users complete more tasks within their usage limits.

The two models serve different purposes. GPT-6.2 Astra is OpenAI’s new frontier model for demanding projects, while GPT-6.2 Sol is designed for everyday work, offering a balance of intelligence, speed and efficiency.

What is new with GPT-6.2 Astra and Sol?

GPT-6.2 Astra focuses on complex professional assignments, including creating financial spreadsheets, preparing legal documents, designing presentations and developing software.

OpenAI says the model produces more polished documents and presentations, follows existing templates more accurately and writes cleaner code.

The company also reported that Astra scored 5.2 points higher than GPT-6 Astra on its Agents’ Last Exam benchmark while completing tasks at less than one-third of the cost.

GPT-6.2 Sol, meanwhile, is positioned as the model for everyday work. It can handle multi-step tasks, write and review code, assist with research and support general workplace activities.

OpenAI says Sol performs close to Astra on some coding evaluations while costing approximately one-fifth as much to run.

Both models also bring improvements to computer use, allowing AI agents to interact more effectively with software, websites and design tools.

Alongside the launch, OpenAI announced changes to API pricing. GPT-6.2 Astra costs $8 per million input tokens, $0.40 per million cached input tokens and $40 per million output tokens. These represent a 20% reduction in standard input and output pricing compared with GPT-6 Astra, alongside a 60% discount on cached input.

GPT-6.2 Sol retains the previous GPT-6.1 Sol pricing of $2 per million input tokens and $10 per million output tokens. The models are available to Plus, Pro, Enterprise and Edu users through ChatGPT Work and Codex. Developers can also access both models through the OpenAI API.

Leave a Reply