🔍 Read the full analysis: GPT-6.2 Astra And Sol: What Their Launch Means For Work on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI released GPT-6.2 Astra for demanding projects and GPT-6.2 Sol for lower-cost everyday work on Nov. 17, 2026, alongside a legal configuration. Independent tests put Astra near the top of a leading model index and show lower cost per tested task than several rivals, but report no improvement in the Coding Agent Index. OpenAI rates both models at its Critical cybersecurity threshold; access details, ChatGPT availability and the practical differences between Astra and Astra for Law remain unclear.
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on Nov. 17, 2026, presenting them as a high-capability model for demanding projects and a lower-cost option for everyday work. The launch also includes Astra for Law 6.2. Independent tests put Astra close to the leading model family on one major index and report lower cost per task than several rivals, while OpenAI’s same-day safety report rates both new models at its Critical cybersecurity threshold.
OpenAI describes Astra as its model for the hardest projects and Sol as its option for everyday work, including coding and running agents at scale. Both models accept text and images, have a 1,050,000-token context window, and can return up to 128,000 output tokens. OpenAI lists Astra’s knowledge cutoff as May 18, 2026, and Sol’s as June 1, 2026. Fine-tuning is not supported.
At standard short-context API rates, Astra costs $8 per million input tokens and $40 per million output tokens; Sol costs $2 and $10, respectively. Astra’s input and output prices are 20% below those of GPT-6 Astra, while Sol retains the earlier $2 and $10 rates. These are list prices, not a measure of what every workload will cost: requests exceeding 272,000 input tokens have higher rates, and speed tiers, batch processing and eligible background requests change charges.
Artificial Analysis’s Intelligence Index, accessed on launch day, gives Astra a score of 62 and Sol 57. Astra’s unrounded score of 62.14 puts it just below Claude Fable 5.2’s 62.22 at the compared settings; its model page ranks Astra Max third among 91 configurations. The index reports a task cost of $4.00 for Astra Max and $1.01 for Sol Max, compared with $26.54 for Claude Fable 5.2 Max and $5.48 for Gemini 4 Carbon High. Artificial Analysis also reports no improvement in its Coding Agent Index and hallucination rates of 59% for Astra and 55% for Sol at maximum effort.
GPT-6.2 Astra, Sol and Astra for Law
Three models, three jobs
What launched, what it costs and how it scores in independent tests. Astra and Sol at their Max setting; the better value in each row is outlined.
- Three releases, one dayOn November 17, OpenAI released GPT-6.2 Astra for the hardest projects and GPT-6.2 Sol for everyday work. Astra for Law 6.2 is a legal configuration at Astra’s price.
- The independent indexAstra scores 62, second among model families behind Claude Fable 5.2 at 65. Sol scores 57. One index task costs $4.00 and $1.01.
- Speed and a legal testAstra needs 10.4 minutes per task, Sol 7.9. On the Harvey LAB-AA legal test, Astra for Law passes 20.8% of tasks.
- OpenAI’s own safety ratingBoth models meet OpenAI’s Critical threshold for cybersecurity. Higher-risk cyber help now needs extra vetting.
Source: Artificial Analysis, model pages for GPT-6.2 Astra, GPT-6.2 Sol and Astra for Law 6.2, accessed Nov 17, 2026 (Intelligence Index, cost and time per task, Harvey LAB-AA v1.1); OpenAI, API pricing page and GPT-6.2 System Card, Nov 17, 2026. Prices per million input/output tokens. Redrawn by Thorsten Meyer AI.
Graphic: Powered by Thorsten Meyer AI
The Cost of Work at Scale
The launch gives businesses and developers another choice between higher capability and lower unit cost, with Sol positioned as a cheaper option for work that may be repeated across many requests. The index’s task-cost figures suggest that, on its tested tasks and configurations, the models can cost less than several named competitors. Those results do not establish savings for every company: actual bills depend on prompt length, output volume, speed tier, account limits and the nature of the work.
For teams considering automated coding, research or agent workflows, the mixed benchmark picture matters as much as the headline ranking. Astra scores strongly overall but did not improve on the Coding Agent Index, and the reported hallucination rates remain relevant for work requiring reliable answers. The Critical cybersecurity rating also has practical consequences: OpenAI says more permissive cyber access now requires extra vetting. Organizations will need to weigh price and capability against testing, access controls and review of model outputs.
Products Worth Considering

Jackery Explorer 300 Portable Power Station, 292Wh

Bose Smart Ultra Dolby Atmos Soundbar with Voice Control, HDMI eARC, Black

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
As an affiliate, we earn on qualifying purchases.
A Two-Model Work Split
OpenAI’s developer guide summarizes the intended division: use Astra for the most demanding tasks or Sol for complex work at lower cost. This is a product positioning, not independent proof that either model is best for a particular job. Artificial Analysis reports that Astra gains five index points over GPT-6 Astra, while Sol gains two points and reaches 57, on par with GPT-6 Astra in that index at about 15% of its cost per task.
OpenAI also listed Astra for Law 6.2 as a legal model configuration with the same standard API prices as Astra. The available model pages do not explain how it differs from the general model. Artificial Analysis had not assigned it an Intelligence Index score on Nov. 17, though it had tested the legal configuration on one legal benchmark.
There are implementation details for developers. OpenAI says the API drops earlier reasoning when conversation history is edited mid-conversation, which can reduce answer quality and cause the model to repeat work using more tokens. Requests setting reasoning effort to “none” are rejected. The listed account tiers start at 5,000 requests and 1 million tokens per minute, with higher tiers reaching 15,000 requests and 40 million tokens per minute.
Questions for Buyers and Developers
Several details were not established in the material available on launch day. OpenAI’s launch post could not be accessed, so which ChatGPT plans include Astra and Sol is not confirmed here. The listed model pages also do not specify how Astra for Law differs from Astra, and its position cannot be compared on the Intelligence Index because it had no score there at the time.
Benchmark scores and task costs describe Artificial Analysis’s tests, not every workplace or deployment. The source material does not give enough detail to infer performance on all coding, legal or office tasks, nor does it establish a typical customer’s savings. The reported hallucination rates are tied to maximum-effort settings; the available information does not specify how they vary across other settings or real-world use.
OpenAI’s system card rates both models at Critical for cybersecurity, the higher of its two risk thresholds, and says more permissive cyber access requires additional vetting. Astra is also described as the first externally released OpenAI model to meet the company’s High threshold for AI self-improvement. These are OpenAI’s safety assessments; the supplied material does not provide enough information to independently validate those ratings or detail every access restriction.
Availability, Pricing and Testing
Developers can check OpenAI’s API model pages and account-specific limits for current access and rates. The launch-day information says approved organizations can use Zero Data Retention and US or EU data residency with Standard, Fast and Batch processing; Ultrafast processing is limited to US data residency. Promotional pricing for Astra for Law is listed as available at least through Dec. 31, 2026. Ultrafast pricing is listed at three times the standard rate through Feb. 28, 2027, and six times the standard rate beginning March 1, 2027.
The next useful evidence will be broader independent testing, especially for coding agents and legal work, and clearer documentation of Astra for Law and ChatGPT-plan availability. Organizations assessing the models can compare them on their own tasks and include token use, latency, review needs and access requirements in the cost calculation. Those details remain open in the launch-day information; the release alone does not establish how much work the models will handle reliably or what they will cost in production.
Key Questions
What did OpenAI launch?
OpenAI released GPT-6.2 Astra for demanding projects and GPT-6.2 Sol for everyday work on Nov. 17, 2026. It also listed Astra for Law 6.2 as a legal configuration.
Which model is cheaper?
At standard short-context API rates, Sol costs $2 per million input tokens and $10 per million output tokens. Astra costs $8 and $40. Longer prompts and speed or processing options can change the bill.
Did Astra lead the independent model index?
No. Artificial Analysis scored Astra Max at 62.14 and Claude Fable 5.2 Xhigh at 62.22. Astra’s model page ranked it third among 91 configurations; the scores were close, but Astra did not lead.
What safety rating did the new models receive?
OpenAI’s system card rates both Astra and Sol at its Critical cybersecurity threshold and says more permissive cyber access requires extra vetting. This is OpenAI’s assessment, not an independent safety verdict.
Is Astra for Law available in ChatGPT?
The launch-day material lists Astra for Law 6.2 as a legal API configuration, but does not say which ChatGPT plans include Astra or Sol or explain the legal model’s differences from Astra.
Source: ThorstenMeyerAI.com
