OpenAI has officially launched GPT-6.2 Astra and GPT-6.2 Sol, two new language models designed for enterprise professional workflows and everyday consumer applications, respectively. Both models are immediately available in ChatGPT Work, Codex, and via the OpenAI API. OpenAI reduced API pricing for GPT-6.2 Astra by 20% compared to its predecessor to $8.00 per million input tokens and $40.00 per million output tokens, while maintaining GPT-6.2 Sol’s pricing at $2.00 per million input tokens.
According to OpenAI, GPT-6.2 Astra delivers performance gains across multi-step enterprise workflows, outperforming competing models on the Agents’ Last Exam benchmark and legal analysis evaluations like Harvey LAB-AA. The flagship model also demonstrates updated capabilities in terminal execution, 3D web design generation, physical assembly modeling, and automated cybersecurity vulnerability detection and patching.
The launch underscores OpenAI’s ongoing push to increase the price-to-performance ratio for enterprise deployments, offering lower latencies and token usage footprints across complex legal, financial, and coding applications.
Why it matters
Developers gain immediate access to lower API pricing for enterprise-grade reasoning and agentic computer-use capabilities.
Improved token efficiency in automated code execution and design tasks directly reduces operational costs for AI-native software products.
Source: the-decoder.com



