OpenAI Launches GPT-6.2 Astra and Sol, Claims Major Leap in AI Performance and Efficiency


OpenAI has officially launched GPT-6.2 Astra and GPT-6.2 Sol, two new models targeting professional workloads and everyday tasks. Both promise better results, lower token consumption, and improved efficiency. Interestingly, OpenAI says Astra offers stronger performance at significantly lower costs than its predecessor.

GPT-6.2 Astra raises the bar for professional AI work

OpenAI describes GPT-6.2 Astra as its new frontier model for complex professional projects. The model focuses on producing better presentations, spreadsheets, documents, and code. It can also follow supplied templates more accurately and create improved website designs.

According to OpenAI, Astra scores 5.2 points higher than GPT-6 Astra on Agents’ Last Exam. It reportedly achieves this while costing less than one-third as much. The model also sets new highs on scientific research evaluations, including Terminal-Bench-Science 0.1.

Another interesting improvement involves cybersecurity. OpenAI says both new models meet its Critical threshold for cybersecurity capabilities. They also meet the High threshold for biological and chemical capabilities, prompting additional safeguards.

GPT-6.2 Sol targets everyday users

Speaking of the second model, GPT-6.2 Sol is designed for everyday work. OpenAI says it improves intelligence and efficiency compared with GPT-6.1 Sol.

The model handles multistep workflows, coding tasks, and general productivity work. It also performs strongly on everyday evaluations while using significantly less money per task than GPT-6 Astra.

Pricing, availability, and API changes

Both models are available starting today to Plus, Pro, Enterprise, and Edu users through ChatGPT Work and Codex. Developers can also access them through the OpenAI API.

GPT-6.2 Astra costs $8 per million input tokens and $40 per million output tokens. Cached input costs just $0.40. Meanwhile, GPT-6.2 Sol retains its predecessor’s pricing at $2 input and $10 output per million tokens. Moreover, OpenAI is also introducing Ultrafast access for eligible subscribers.

More about the topics: AI, ChatGPT, OpenAI

Readers help support Windows Report. We may get a commission if you buy through our links. Tooltip Icon

Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more

User forum

0 messages