Last updated: November 17, 2026
GPT-6.2 Astra and GPT-6.2 Sol are OpenAI’s new models, released November 17, 2026. Astra is the frontier model for the hardest projects at $8 input and $40 output per million tokens. Sol handles everyday work at $2 and $10. Both have a 1,050,000-token context window and are in ChatGPT, Codex and the API.
Contents
- What is new in GPT-6.2
- GPT-6.2 Astra vs GPT-6.2 Sol
- Pricing, tiers and long-context rates
- Who can use GPT-6.2
- How to migrate an API integration
- Prompting changes to expect
- Limitations and safeguards
- Frequently asked questions
What is new in GPT-6.2
In its announcement, OpenAI describes GPT-6.2 Astra as its new frontier model for demanding projects and GPT-6.2 Sol as an efficient, versatile model for everyday work. The common theme is efficiency: OpenAI says both models deliver better results on demanding tasks with fewer output tokens.
For Astra, OpenAI highlights three areas:
- Professional documents. Slides, spreadsheets and documents that follow your templates and the conventions of fields such as law, finance and accounting. OpenAI says Astra can match a slide template from a file or a screenshot.
- Code and frontends. Cleaner code style, more focused tests and better implementation of complex changes in existing codebases.
- Computer use. More precise control, down to individual parts in CAD and simulation tools, and better judgment about when to use plugins instead of computer use.
OpenAI’s reported results include a 6.2-point gain over GPT-6 Astra on its internal GDPval v2 benchmark while using 22% fewer output tokens, and a 5.3-point gain for GPT-6.2 Sol over GPT-6.1 Sol. On Cognition’s FrontierCode 2.0, OpenAI says Sol came within 0.5 points of Astra at about one-fifth the cost. These are vendor-reported figures.
GPT-6.2 Astra vs GPT-6.2 Sol
The specifications below come from the GPT-6.2 Astra model page and the GPT-6.2 Sol model page.
| Specification | GPT-6.2 Astra | GPT-6.2 Sol |
|---|---|---|
| API model ID | gpt-6.2-astra |
gpt-6.2-sol |
| Best for | Hardest projects, start to finish | Everyday work and agents at scale |
| Input / cached input / output (per 1M tokens) | $8.00 / $0.40 / $40.00 | $2.00 / $0.10 / $10.00 |
| Cache write, 30 minutes (per 1M) | $10.00 | $2.50 |
| Context window | 1,050,000 tokens | 1,050,000 tokens |
| Max input / max output | 922,000 / 128,000 tokens | 922,000 / 128,000 tokens |
| Knowledge cutoff | May 18, 2026 | June 1, 2026 |
| Inputs / outputs | Text and image in, text out | Text and image in, text out |
| Change from previous version | Input and output 20% cheaper than GPT-6 Astra; cached input 60% cheaper | Same rates as GPT-6.1 Sol |
Pick Astra for long, high-stakes deliverables and complex code changes. Pick Sol as a default for coding, writing and multistep agents, where OpenAI says it lets you finish more tasks on the same paid plan than GPT-6.1 Sol. For the previous lineup, see our GPT-6.1 Terra guide.
Pricing, tiers and long-context rates
The model pages list several pricing rules that matter for heavy users:
- Long context. For requests over 272K input tokens, input and cache rates double and output costs 1.5 times the short-context rate for the full request.
- Longer cache. A 12-hour cache write costs $20 per million tokens on Astra and $5 on Sol for short-context requests at Standard pricing.
- Batch costs 50% of Standard. Fast mode costs twice Standard.
- Ultrafast costs 3 times Standard through February 28, 2027, and 6 times Standard from March 1, 2027 (Pacific time).
The API changelog also lists a new public beta in the Responses API: a two-hour or eight-hour background completion window, with token rates 30% and 40% below Standard. It covers GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Terra and GPT-6 Luna. For wider context, see our OpenAI API pricing explainer.
Who can use GPT-6.2
According to OpenAI, both models are available from launch day to all Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex. Enterprise and Edu users get access once their workspace admin enables it.
Ultrafast differs by plan. Pro $500 subscribers can use Ultrafast with either model. Pro $200 subscribers get Ultrafast for the first time, starting with GPT-6.2 Sol. In the API, Ultrafast is available for both models at launch. Our ChatGPT Plus vs Pro and Codex usage limits pages explain the plans in more detail.
How to migrate an API integration
OpenAI’s Using GPT-6.2 guide lists the changes to make. In order:
- Set
modeltogpt-6.2-astraorgpt-6.2-sol. - Set reasoning effort. Start with
mediumfor a new integration. If you usenoneorminimaltoday, switch tolow; the API rejectsnone. - Move tool calling to the Responses API. Chat Completions works, but tool calling requires Responses.
- Remove unsupported sampling parameters:
temperature,top_pandtop_logprobs, pluslogprobsin Chat Completions. - If you are coming from GPT-5.5 or earlier, replace
prompt_cache_retentionwithprompt_cache_options.ttlset to"30m". - Keep the prompt prefix unchanged and append new instructions instead of editing earlier turns.
The last step matters more than before. Starting with GPT-6.2, the API drops reasoning items when it detects mid-conversation edits to the history, a model downgrade, or a change of end user or organization. OpenAI warns that answer quality can decline and the model may spend more tokens repeating work.
Other details from the guide: both models support Zero Data Retention for approved organizations, US and EU data residency with Standard, Fast and Batch, and US-only data residency for Ultrafast.
Prompting changes to expect
The guide notes four behaviors of GPT-6.2 Astra:
- Writing is more concise and direct by default, so revisit instructions like “be concise”.
- Task completion. When intent is unclear, the model may propose a next step and wait for confirmation. For long tasks, tell it which decisions it can make on its own.
- Coding style. Describe project conventions you want it to follow.
- Frontend design. Specify the imagery and typography a page should use.
Limitations and safeguards
- Tighter cyber safeguards. OpenAI says both models meet its Critical threshold for cybersecurity and launch with restrictions that also limit some legitimate security testing. GPT-6.2 is not yet on Daybreak.
- Disclosure of obstacles. OpenAI reports that in two tests the models were more likely to omit obstacles such as a missing file. It added Codex instructions to reduce this.
- No audio or video. The model pages list audio and video as not supported, and fine-tuning as not supported.
- Benchmarks are vendor-reported. Compare results on your own tasks before switching.
For security research access, see our GPT-6 Cyber guide. For the closest competitor, see our Claude Fable 5.2 guide.
Frequently asked questions
When was GPT-6.2 released?
OpenAI announced GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026. Both models became available the same day in ChatGPT Work, Codex and the OpenAI API.
How much does GPT-6.2 Astra cost in the API?
Standard rates are $8 per million input tokens, $0.40 per million cached input tokens and $40 per million output tokens. A 30-minute cache write costs $10 per million tokens. Requests over 272K input tokens cost more.
What is the difference between GPT-6.2 Astra and GPT-6.2 Sol?
Astra is OpenAI’s frontier model for the hardest projects. Sol is the efficient model for everyday coding, writing and agent workflows. Sol costs a quarter of Astra’s per-token rate and has the same 1,050,000-token context window.
Who can use GPT-6.2 in ChatGPT?
Plus, Pro, Enterprise and Edu users can use both models in ChatGPT Work and Codex. Enterprise and Edu users get access once a workspace admin enables it. Pro $200 subscribers get Ultrafast with Sol, and Pro $500 subscribers get Ultrafast with both.
Does GPT-6.2 support temperature?
No. OpenAI’s migration guide says to omit temperature, top_p and top_logprobs from Responses requests, and also logprobs from Chat Completions requests. The API also rejects a reasoning effort of none.
Sources
- OpenAI: Introducing GPT-6.2 Astra and Sol
- OpenAI API docs: Using GPT-6.2
- OpenAI API docs: GPT-6.2 Astra model page
- OpenAI API docs: GPT-6.2 Sol model page
- OpenAI API docs: Changelog, November 17, 2026
Image: GPT-6.2 Astra and Sol announcement image from OpenAI, used for editorial purposes.
Related guide: OpenAI background mode guide to 2h and 8h completion windows.
1 thought on “GPT-6.2 Astra and Sol: Power-User Guide to Pricing, API and Migration”