GPT-6.2 Astra and Sol: Pricing, API Access and What Changed (2026)

OpenAI released two new models on November 17, 2026: GPT-6.2 Astra and GPT-6.2 Sol. Astra is the flagship for hard professional work. Sol is the cheaper model for everyday tasks. The headline change is price: Astra’s API rates drop 20% compared with GPT-6 Astra, while Sol keeps the same rates as GPT-6.1 Sol. Here is what is confirmed and how to choose between them.

What OpenAI Released

GPT-6.2 Astra is positioned for complex projects where quality matters most. GPT-6.2 Sol is positioned for daily work such as drafting, coding help and quick analysis. OpenAI says both models deliver better results while using fewer output tokens, which is where much of the cost saving comes from in real use.

Benchmark numbers in the launch come from OpenAI and its partners, so treat them as vendor claims until independent testing confirms them in daily work.

GPT-6.2 API Pricing

Standard-speed API prices per million tokens:

ModelInputCached inputOutput
GPT-6.2 Astra (gpt-6.2-astra)$8$0.40$40
GPT-6.2 Sol (gpt-6.2-sol)$2$0.10$10
GPT-6 Astra (previous)$10$1.00$50

The cached input rate for Astra falls sharply, from $1.00 to $0.40 per million tokens. If your app resends the same long instructions or documents, caching is the biggest lever on your bill.

Where You Can Use It

  • ChatGPT Work and Codex, for paid subscribers on Plus, Pro, Enterprise and Edu tiers, with reports also listing Business.
  • The OpenAI API, using the model IDs gpt-6.2-astra and gpt-6.2-sol, with Responses and Chat Completions support.
  • An Ultrafast mode is being offered for the Pro 200 plan.

Both models support up to 128K tokens of output, according to published model details.

Astra vs Sol: Which One Should You Pick?

Start with Sol for everyday coding, writing and support tasks. It costs a fifth of Astra on input and output. OpenAI’s own numbers show Sol landing within half a point of Astra on the FrontierCode 2.0 test at about a fifth of the cost, which is a strong argument for defaulting to Sol and escalating only when needed.

Move to Astra for the hardest jobs: long multi-step analysis, difficult debugging, and professional documents where one wrong detail is expensive.

How It Compares With Claude

Astra’s new $8 input and $40 output rate matches the list price of Anthropic’s Claude Fable 5.2, which launched on November 9. That moves the competition away from list price and toward cost per finished task, meaning how many tokens a model needs to get the job right. For more on the Anthropic side, read our guide to Claude Fable 5.2 price, speed and the Anthropic IPO.

What To Do Now

  1. If you use gpt-6-astra in the API, test gpt-6.2-astra on a sample of real prompts and compare both quality and total tokens used.
  2. If you run high-volume workloads, trial gpt-6.2-sol first and measure how often it needs a retry.
  3. Enable prompt caching for repeated context to benefit from the lower cached rate.
  4. Do not switch production traffic on benchmark claims alone. Run your own evaluation.

Frequently Asked Questions

When was GPT-6.2 released?

OpenAI published GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026.

Is GPT-6.2 Astra cheaper than GPT-6 Astra?

Yes. Standard API prices are $8 input and $40 output per million tokens, about 20% below GPT-6 Astra.

Did Sol get cheaper?

No. Sol keeps the $2 input and $10 output rates of GPT-6.1 Sol. The gains are in speed and shorter answers.

Which ChatGPT plans include GPT-6.2?

Paid plans including Plus, Pro, Enterprise and Edu can choose either model in ChatGPT Work or Codex.

Is Astra worth the extra cost?

Only for the hardest work. For most daily tasks Sol delivers close results at a much lower price.

Pricing and availability can change, so confirm current rates on OpenAI’s official pricing page before you budget. Follow AI Trends Setter for more daily AI model updates.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest articles

Related articles