OpenAI has launched GPT-6.2 Astra and GPT-6.2 Sol, two updated artificial intelligence models aimed at professional workloads and software development. Starting today, both models are accessible in the API, as well as in ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu subscribers. They are not available on free tiers or in standard Chat mode. For API developers, Astra costs $8 per million input tokens, $0.40 for cached input, and $40 for output, representing a 20% drop in standard pricing compared to GPT-6 Astra. Sol maintains its previous pricing of $2 per million input tokens, $0.10 for cached input, and $10 for output. [2][3][4]
OpenAI positions Astra as its premium option for complex documents, presentations, and coding tasks, while Sol serves as an efficient tool for everyday use. OpenAI says both models achieve better results while using fewer output tokens to help users maximize plan allowances. Additionally, subscribers paying $200 per month for ChatGPT Pro now gain access to Ultrafast mode when using Sol. Previously restricted to $500 monthly Pro plans, Ultrafast mode is also open to API developers, as well as Enterprise and Edu workspaces once administrators enable it. [2][3][4]
The release comes a week after Anthropic introduced Claude Fable 5.2 and Google announced Gemini 4 Hydrogen. OpenAI claims that on the Agents' Last Exam benchmark, Astra beat Claude Fable 5.2 by 5.5 points at 5% of the estimated cost. On Cognition's FrontierCode 2.0 evaluation, OpenAI says Sol came within half a point of Astra's score at roughly one-fifth the expense. [2][3][4]