AI & Innovation

GPT-6.2 Astra and Sol: OpenAI's new models, prices and who gets them

By Anika SharmaAI & Innovations Editor, Peepals Tech
WhatsAppX
Representative image: person typing on laptop with chatbot

Representative image: kaboompics / Pixabay

OpenAI released GPT-6.2 Astra for professional work and GPT-6.2 Sol for everyday tasks on 17 November 2026, cutting Astra's API price by 20%.

GPT 6.2: what happened

OpenAI released two new artificial intelligence (AI) models, GPT-6.2 Astra and GPT-6.2 Sol, on Tuesday, 17 November 2026. Astra is the higher-end model, aimed at demanding professional work; Sol is a cheaper model for everyday tasks. Both are available in ChatGPT Work, in Codex, OpenAI's coding tool, and through its API, the paid interface developers use to build the models into their own products. OpenAI cut Astra's standard API price by 20% compared with the earlier GPT-6 Astra, while Sol keeps the price of the model it replaces.

GPT 6.2: key facts

  • Launch: Tuesday, 17 November 2026, announced on OpenAI's developer community forum.
  • GPT-6.2 Astra: for professional work; API price $8 per million input tokens and $40 per million output tokens, 20% less than GPT-6 Astra.
  • GPT-6.2 Sol: for everyday work; $2 per million input tokens and $10 per million output tokens, the same as GPT-6.1 Sol.
  • Where: ChatGPT Work, Codex and the API.
  • Who: Plus, Pro, Enterprise and Edu users; Enterprise and Edu workspaces need an administrator to switch the models on.
  • OpenAI's claims: Astra scores 5.2 points higher than GPT-6 Astra on the Agents' Last Exam benchmark at under a third of the cost, and 6.2 points higher on its internal GDPval v2 test while using 22% fewer output tokens.
  • Safety: OpenAI has rated Astra "High" for AI self-improvement in its safety framework, the first externally released OpenAI model with that rating, and added tighter cybersecurity safeguards.

GPT 6.2: why it matters

A "token" is a small chunk of text, roughly a part of a word, and AI companies charge developers by the million tokens a model reads (input) and writes (output). Price cuts at the top end matter because many businesses, including Indian software firms building AI agents for clients, pay per token at scale. A 20% cut on the most capable model can decide whether a product that runs thousands of AI calls a day makes money.

The two-tier launch also shows how AI companies now sell models: one expensive model for hard, multi-step work such as research, legal drafting or long coding tasks, and one much cheaper model for routine requests. Independent testers have already started checking OpenAI's claims. One independent benchmarking firm, Artificial Analysis, gave Sol a score that matches the original Astra at about 15% of its cost per task, while another independent report found no improvement for Astra as a coding agent. As always with new models, the company's benchmark numbers are its own and should be read alongside outside tests.

The safety rating deserves attention. "AI self-improvement" refers to a model's ability to help speed up AI research itself, including work on future AI systems. OpenAI paired the release with tighter cybersecurity safeguards. It is the company's assessment, not an outside regulator's, but it signals that frontier models are reaching levels their makers treat as sensitive.

For Indian users, the models arrive through the same paid ChatGPT plans and API as elsewhere; OpenAI did not announce India-specific pricing.

Peepals calculation: a 20% cut that brings Astra to $8 and $40 means the earlier GPT-6 Astra cost $10 per million input tokens and $50 per million output tokens (8 ÷ 0.8 = 10; 40 ÷ 0.8 = 50). Sol, at $2 and $10, now costs one-quarter of Astra on both input and output (2 ÷ 8 = 0.25; 10 ÷ 40 = 0.25). Our calculation from the reported figures.

ModelInput price (per million tokens)Output price (per million tokens)
GPT-6.2 Astra$8$40
GPT-6 Astra (earlier)$10$50
GPT-6.2 Sol$2$10

GPT 6.2: impact

Short-term: Paid ChatGPT users and developers can switch to the new models now; companies using Astra through the API pay less for the same volume of work.

Long-term: Cheaper top-end models make AI agents, which carry out multi-step tasks on their own, more affordable to run, adding pressure on rival model makers to cut prices too.

Who is affected: ChatGPT Plus, Pro, Enterprise and Edu subscribers; developers and start-ups building on OpenAI's API, including Indian IT and SaaS firms; and rival AI companies competing on price.

GPT 6.2: key takeaway

Developers paying for GPT-6 Astra should test GPT-6.2 Astra and the cheaper Sol on their own tasks, since independent results so far differ from OpenAI's own benchmark claims.

GPT 6.2: questions and answers

What is GPT 6.2?

GPT-6.2 is OpenAI's model family released on 17 November 2026, in two versions: Astra for demanding professional work and Sol for everyday tasks.

How much does GPT-6.2 Astra cost in the API?

$8 per million input tokens and $40 per million output tokens, 20% less than GPT-6 Astra.

Can free ChatGPT users use GPT-6.2 Astra?

OpenAI listed Plus, Pro, Enterprise and Edu users as getting the models in ChatGPT Work and Codex; it did not list the free tier.

What does the "High" self-improvement rating mean?

It is OpenAI's own safety-framework rating for a model's ability to help advance AI research; Astra is the first externally released OpenAI model to get it, and OpenAI added tighter cybersecurity safeguards with the launch.

Disclaimer: Prepared by the Peepals newsroom from publicly available sources with AI assistance. Information is accurate to the best of our knowledge at the time of publication and may change. Images may be representative. Not professional advice. Report an error via our contact page.

#GPT 6.2#GPT-6.2 Astra#GPT-6.2 Sol#OpenAI#ChatGPT

By Anika Sharma, AI & Innovations EditorCovers AI & Innovation and more for Peepals Tech. All stories by Anika Sharma

Your week ahead, every Sunday. Free.

Exam and result dates, IPOs opening, scheme deadlines, government holidays and the RBI decision for the next 7 days, plus the week's top stories. One email a week, unsubscribe in one click.

Anika Sharma is a pen name of the Peepals Global Editorial Team, which researches, fact-checks and publishes every story. The portrait is an illustration.

Responsible for this story: the Peepals Global Editorial Team, led by editor Prashant Sonar. How we report

More from Peepals Tech

Your week ahead, every Sunday. Free.

Exam and result dates, IPOs opening, scheme deadlines, government holidays and the RBI decision for the next 7 days, plus the week's top stories. One email a week, unsubscribe in one click.