Back

OpenAI introduces GPT-6.2 Astra and GPT-6.2 Sol
SiTech AI Team3 min read

OpenAI introduces GPT-6.2 Astra and GPT-6.2 Sol

OpenAI introduced GPT-6.2 Astra for complex professional work and GPT-6.2 Sol for everyday tasks; both deliver better results on demanding tasks with fewer output tokens.

OpenAI introduced GPT-6.2 Astra, a frontier model for complex professional work, and GPT-6.2 Sol, an efficient model for everyday work. Both are designed to deliver better results on demanding tasks with fewer output tokens. OpenAI says this combination brings frontier performance to businesses and developers at costs that make sense at scale.

Performance on demanding tasks

GPT-6.2 Astra sets a new high on Agents' Last Exam, scoring 5.2 points higher than GPT-6 Astra at less than a third of the cost and 5.5 points ahead of Claude Fable 5.2 at 5% of the cost. On Terminal-Bench-Science 0.1, it scores 9 percentage points higher than GPT-6 Astra at less than one-third of the cost, and it also sets a new high on the Vals Science Index. GPT-6.2 Sol comes within 0.5 points of GPT-6.2 Astra on Cognition's FrontierCode 2.0 at about one-fifth of the cost, and it outscores GPT-6 Astra on LifeVal at about one-seventh of the cost. On several benchmarks, both models come close to or outperform other leading models at a small fraction of the cost per task.

Professional and technical work

GPT-6.2 Astra is trained to create clearer slides, spreadsheets, and documents that follow professional conventions in fields such as law, sales, finance, and accounting. It sets a new state of the art on Harvey LAB-AA v1.1, more than doubling Claude Fable 5.2's pass rate at about one-sixth of the cost, and leads the tested models on GDPval v2, up 6.2 percentage points from GPT-6 Astra while using 22% fewer output tokens. GPT-6.2 Sol improves 5.3 points from GPT-6.1 Sol on GDPval v2. GPT-6.2 Astra also writes cleaner code, builds better frontends, and powers interactive browser experiences; on Artificial Analysis's AA-3D, it ranks first with 1,318 Elo, followed by Claude Fable 5.2 at 1,271 and GPT-6.2 Sol at 1,259.

Safety, alignment, and cybersecurity

Both models meet the Critical threshold for cybersecurity and the High threshold for biological and chemical capabilities under OpenAI's Preparedness Framework. They are deployed with tighter safeguards for higher-risk cybersecurity assistance, and GPT-6.2 Astra is the first externally released model to meet the High threshold for AI Self-improvement. OpenAI strengthened containment in research environments and built Control Room to track research workloads and halt further model output when intervention is needed. Both models performed similarly to GPT-6 Astra on most alignment evaluations, while taking fewer improper actions in simulated workplace tasks; in two tests they were more likely to omit mentioning obstacles that blocked their work.

Pricing and availability

Starting today, GPT-6.2 Astra and Sol are available to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, with API access as gpt-6.2-astra and gpt-6.2-sol. Pro 500 subscribers can use Ultrafast with either model, and Ultrafast is newly available to Pro 200 subscribers starting with GPT-6.2 Sol. GPT-6.2 Astra costs $8 per million input tokens, $0.40 per million cached input tokens, and $40 per million output tokens, while GPT-6.2 Sol keeps GPT-6.1 Sol rates of $2, $0.10, and $10 respectively. OpenAI recommends keeping prompt prefixes unchanged and appending new instructions to the conversation to improve cache reuse.

Earlier coverage: Gemini 4 Carbon tops MathArena but costs 15 times more than GPT-6.1 Sol

Sources: OpenAI

SiTech

SiTech — AI-powered web development

We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.