OpenAI Releases Two GPT-6.2 Models and Lowers API Prices

2026-11-18 10:13
Favorite

en.Wedoany.com Reported - On November 17, the U.S. artificial intelligence company OpenAI officially released two models, GPT-6.2 Astra and GPT-6.2 Sol, targeting complex professional tasks and everyday general work, respectively. GPT-6.2 Astra improves processing capabilities in tasks such as document creation, software development, and scientific research, with API input and output token prices reduced by 20% compared with the previous generation GPT-6 Astra, and cached input prices reduced by 60%. Both models are now available through ChatGPT Work, Codex, and the OpenAI API.

GPT-6.2 Astra focuses on upgrading capabilities in presentation decks, spreadsheets, document generation, front-end development, and scientific research task processing. According to test results published by OpenAI, the model's score on the multi-step professional task benchmark Agents' Last Exam is 5.2 percentage points higher than that of GPT-6 Astra, while the cost of completing tasks has dropped to less than one-third of the previous generation. On the scientific research benchmark Terminal-Bench-Science 0.1, the new model's score is 9 percentage points higher than the previous generation, with task costs also falling to less than one-third.

GPT-6.2 Sol targets everyday office work, programming, and multi-step agent workflows. OpenAI's published FrontierCode 2.0 test results show that the score gap between this model and GPT-6.2 Astra is approximately 0.5 percentage points, with task costs about one-fifth of the latter. Sol continues to use the API pricing of GPT-6.1 Sol, charging $2 per million input tokens, $10 per million output tokens, and $0.10 for cached input.

GPT-6.2 Astra's standard API pricing has been adjusted to $8 per million input tokens, $0.40 per million cached input tokens, and $40 per million output tokens. Both models support the Responses API and Chat Completions interface, and developers can invoke them via gpt-6.2-astra and gpt-6.2-sol, respectively, with support for batch and flexible processing services.

Both models are now available to Plus, Pro, Enterprise, and Edu users of ChatGPT Work and Codex. Pro 500 users can use the Ultrafast high-speed mode of both models, and Pro 200 users gain access to Ultrafast for GPT-6.2 Sol for the first time. Eligible Enterprise and Edu users can access the relevant models after workspace administrators enable them.

The GPT-6.2 system card released alongside the models shows that both models reach the "Critical" level for cybersecurity capabilities and the "High" level for biological and chemical capabilities under OpenAI's Preparedness Framework. Among them, GPT-6.2 Astra reaches the "High" level for AI self-improvement capabilities, becoming OpenAI's first publicly released model to achieve this level. The company has strengthened default restrictions on high-risk cybersecurity assistance and set additional review requirements for higher-privilege use.

This bulletin is compiled and reposted from information of global Internet and strategic partners, aiming to provide communication for readers. If there is any infringement or other issues, please inform us in time. We will make modifications or deletions accordingly. Unauthorized reproduction of this article is strictly prohibited. Email: news@wedoany.com