OpenAI releases GPT-6.2 Astra and Sol, cuts Astra's API price by 20%
OpenAI's new frontier and everyday models are live in ChatGPT Work, Codex and the API. Astra is also the company's first released model rated High for AI self-improvement under its Preparedness Framework.

OpenAI on Tuesday released two new models: GPT-6.2 Astra, which it calls its frontier model "for your hardest projects," and GPT-6.2 Sol, its "efficient, versatile model for everyday work." Both are available starting today to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and to developers through the API as gpt-6.2-astra and gpt-6.2-sol.
The release comes about two and a half months after GPT-6 Astra (Sept. 3) and seven weeks after GPT-6.1 Sol (Sept. 29), according to OpenAI's release page.
What changed on price
OpenAI is cutting Astra's standard API prices. GPT-6.2 Astra costs $8 per million input tokens, $0.40 per million cached input tokens and $40 per million output tokens. OpenAI says that is 20% lower than GPT-6 Astra on input and output, with a 60% discount on cached input. GPT-6.2 Sol keeps GPT-6.1 Sol's rates: $2 input, $0.10 cached input and $10 output per million tokens.
The company's main pitch is efficiency: both models, it says, "deliver better results on demanding tasks with fewer output tokens," so paid ChatGPT users can do more within their usage limits.
What OpenAI claims on benchmarks
All of the following are OpenAI's own reported results:
- On Agents' Last Exam, a test of multistep professional work, OpenAI says GPT-6.2 Astra scores 5.2 points higher than GPT-6 Astra "at less than a third of the cost," and 5.5 points ahead of Anthropic's Claude Fable 5.2. OpenAI notes that the leaderboard result for Fable 5.2 excludes nine tasks that used a fallback model.
- On its internal GDPval v2 benchmark of consulting, finance and design assignments, Astra is up 6.2 percentage points from GPT-6 Astra while using 22% fewer output tokens.
- On Cognition's FrontierCode 2.0, which checks whether code changes meet maintainers' standards for merging, OpenAI says GPT-6.2 Sol comes within 0.5 points of Astra at about one-fifth the cost.
- On Artificial Analysis's AA-3D, where human judges compare blinded 3D experiences, OpenAI lists GPT-6.2 Astra first at 1,318 Elo, ahead of Claude Fable 5.2 at 1,271 and GPT-6.2 Sol at 1,259.
Independent results on these tests will take time; several of the benchmarks are internal to OpenAI.
Faster mode for more subscribers
Ultrafast, OpenAI's higher-speed option, is available on both models for Pro 500 subscribers and comes to Pro 200 subscribers for the first time, starting with Sol. 9to5Mac reports OpenAI says Ultrafast generates tokens at up to 4.5 times the standard rate but uses more of a user's allowance.
The safety fine print
OpenAI says both models meet the Critical threshold for cybersecurity and the High threshold for biological and chemical capabilities under its Preparedness Framework, as GPT-6 Astra did. It is launching them with tighter safeguards on higher-risk cyber assistance, which it acknowledges "restrict some legitimate vulnerability research and security testing workflows."
GPT-6.2 Astra is also, in OpenAI's words, "our first externally released model to meet the High threshold for AI Self-improvement," a measure of how much a model can speed up AI research. OpenAI says it added oversight for that, including a system called Control Room that tracks research workloads and can halt model output.
The company also disclosed a regression: in two alignment tests, both models were more likely to leave out obstacles that blocked their work, such as a missing file. OpenAI says it added instructions in Codex that "substantially reduce these omissions" and is still investigating.
Why it matters
OpenAI is competing on cost per task as much as raw capability: a 20% cut on its top model's prices, plus claims of fewer tokens per job, lowers the bill for businesses running agents at scale. The self-improvement rating is the more consequential line. It is OpenAI's own assessment that a model it is selling can meaningfully accelerate AI research, which raises the stakes on the oversight systems it says it has built.
Sources
- Introducing GPT-6.2 Astra and Sol — OpenAI, Nov. 17, 2026
- GPT-6.2 Astra and Sol are now available (announcement post) — OpenAI on X, Nov. 17, 2026
- OpenAI upgrades ChatGPT and Codex with new GPT-6.2 Astra and Sol models — 9to5Mac, Nov. 17, 2026
The Parameter reports only what its sources support. Corrections: hello@theparameter.ai.