🔍 Read the full analysis: What Do GPT-6.2 Astra And Sol Bring To Work? on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI released GPT-6.2 Astra for demanding work and GPT-6.2 Sol for lower-cost everyday tasks on November 17, 2026. Independent tests place Astra near the top of a major model index and find lower costs per benchmark task than several rivals, but report no improvement in coding-agent performance. OpenAI rates both models Critical for cybersecurity risk.
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026, positioning Astra for demanding projects and Sol for lower-cost everyday work, including coding and agents. Same-day independent testing found strong results and lower costs per benchmark task than several competitors, but no improvement in coding-agent performance; OpenAI’s safety report rated both models Critical for cybersecurity risk.
OpenAI describes Astra as its model for the hardest projects and Sol as its everyday option. The developer guide recommends Astra for the most demanding tasks and Sol for complex work at lower cost. Both models accept text and images and generate text, with a 1,050,000-token context window, up to 922,000 input tokens and up to 128,000 output tokens. OpenAI lists knowledge cutoffs of May 18, 2026, for Astra and June 1 for Sol. Fine-tuning is not supported.
Standard API prices are $8 per million input tokens and $40 per million output tokens for Astra, and $2 and $10, respectively, for Sol. Astra’s rates are 20% below those of the previous GPT-6 Astra, according to the supplied pricing information; Sol retains its listed $2 and $10 rates. For one Artificial Analysis benchmark task, Astra at Max effort cost $4 and Sol at Max cost $1.01. The same task cost $26.54 for Claude Fable 5.2 at Max and $5.48 for Gemini 4 Carbon at High. Those figures apply to that index task, not every workload.
OpenAI also listed GPT-6.2 Astra for Law, a legal model configuration priced at Astra’s standard rates. Its model ID is gpt-6.2-astra-law. Artificial Analysis had not assigned it an Intelligence Index score when the source material was accessed, though it had tested the model on one legal benchmark. OpenAI’s guide also flags an API behavior change: when conversation history is edited mid-conversation, GPT-6.2 drops earlier reasoning, which the company says can reduce answer quality and lead to repeated work and higher token use. Requests setting reasoning effort to “none” are rejected.
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 Complex AI Work
The launch offers developers a clearer choice between higher capability and lower per-task cost. Artificial Analysis reports that Astra scores 62 at Max effort and Sol scores 57 on its Intelligence Index. Claude Fable 5.2 leads the listed model families with 65. Astra’s unrounded score, 62.14, is close to Fable’s 62.22 at Xhigh; the small gap matters because Astra’s model page ranks it third of 91 configurations, even though the rounded family-level comparison can make it appear second.
The cost figures may affect decisions about which model to use for repeated or agent-based tasks. Sol’s reported $1.01 cost for the index task is below Astra’s $4 and the cited costs for the two rival configurations. But benchmark task costs do not establish what a particular business will pay: actual bills depend on prompt length, output, processing tier and request pattern. Prompts above 272,000 input tokens incur higher rates across the whole request. Batch processing is listed at half the Standard rate, while Fast costs twice as much. Ultrafast costs three times Standard through February 28, 2027, then six times from March 1, according to the supplied pricing details.
Performance is not uniformly better. Artificial Analysis reports no improvement in its Coding Agent Index and hallucination rates of 59% for Astra and 55% for Sol at Max effort. These are results from specific evaluations, not a guarantee of behavior on every task. For workplaces, the practical question is whether a model’s performance on their own work justifies its cost and whether human review can manage errors.
Safety is another operational consideration. OpenAI’s system card places both models at its Critical cybersecurity threshold, its higher of two risk thresholds. The report also identifies regressions alongside improvements. That assessment does not by itself establish that either model has caused harm, but it signals that access and deployment controls matter. OpenAI says more permissive cyber access requires extra vetting.
Products Worth Considering

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

Logitech MX Master 4 Ergonomic Wireless Mouse with Haptic Feedback, Graphite

Jackery Explorer 300 Portable Power Station, 292Wh
As an affiliate, we earn on qualifying purchases.
How the Models Are Positioned
The two-model split follows OpenAI’s stated use cases: Astra for its most demanding work and Sol for complex tasks at lower cost. The shared context capacity and input types mean the distinction is not a difference in advertised prompt length or modality. Instead, OpenAI presents the models as different choices for capability and expense. The release also includes a legal configuration, but the available pages do not explain precisely how Astra for Law differs from standard Astra.
Artificial Analysis evaluates models across nine tests and reports scores, task costs and completion times. Its November 17 assessment says OpenAI now trails only Anthropic’s Claude Fable 5.2 at the model-family level. It reports Astra five points above GPT-6 Astra and Sol two points higher, with Sol described as roughly on par with GPT-6 Astra at about 15% of its cost per task. The figures are tied to the index’s tests and configurations; they should not be treated as a universal ranking across every kind of workplace task.
The source article disclosed that it was researched and drafted with AI assistance, including Anthropic’s Claude, and that Anthropic competes with OpenAI. It also said neither OpenAI, Anthropic nor Artificial Analysis reviewed or sponsored the article. The launch post itself was not accessible to the source author, so the available information does not establish which ChatGPT plans include the new models. API pricing and availability details cited here were accessed on November 17, 2026.
Open Questions for Deployments
Several details remain unsettled or absent from the available material. OpenAI’s launch post was not accessible to the source author, so ChatGPT plan availability could not be confirmed. The documentation also does not explain how Astra for Law differs technically from Astra, and the legal configuration had no Intelligence Index score at the time of the assessment.
The reported benchmark and cost results are snapshots from tests accessed on November 17, 2026. They do not show how the models will perform across every company’s software, data or workflows. The source material also does not provide enough detail to evaluate the underlying coding-agent test results or the specific safety regressions. OpenAI’s Critical rating is its own risk assessment; the supplied information does not detail the vetting process or the exact access restrictions that will apply to each user.
Pricing can also vary materially with usage. Rates change for long prompts and speed tiers, and the source lists promotional pricing for Astra for Law through at least December 31, 2026. Account-based rate limits vary by usage tier. Organizations should verify the current pricing, access, data residency and retention terms for their account before estimating deployment costs.
What Buyers Should Check Next
Developers and organizations considering the models can compare them on representative tasks, tracking accuracy, completion time, token use and the need for human review rather than relying on index scores alone. They will also need to account for the API’s revised behavior when conversation history is edited, since OpenAI warns that this can cause repeated work and lower answer quality.
OpenAI’s documentation and pricing pages are the immediate sources to watch for updates on plan availability, Astra for Law’s capabilities, usage limits and promotional terms. More independent testing could clarify performance across coding, legal and other workloads. For deployments involving cybersecurity-sensitive tasks, organizations should account for OpenAI’s Critical ratings and any vetting or access requirements before granting broader use.
Key Questions
What is the difference between GPT-6.2 Astra and Sol?
OpenAI positions Astra for the most demanding tasks and Sol for complex everyday work at lower cost. Its standard API rates are $8 per million input tokens and $40 per million output tokens for Astra, compared with $2 and $10 for Sol.
How did Astra and Sol perform in independent testing?
Artificial Analysis reported Intelligence Index scores of 62 for Astra at Max effort and 57 for Sol at Max. Claude Fable 5.2 led the cited model-family comparison at 65. The evaluator also reported no improvement in the Coding Agent Index and hallucination rates of 59% for Astra and 55% for Sol at Max effort.
Are the new models cheaper than competitors?
In one Artificial Analysis index task, Astra at Max cost $4 and Sol at Max cost $1.01. The same task cost $26.54 for Claude Fable 5.2 at Max and $5.48 for Gemini 4 Carbon at High. These are benchmark-specific costs, not estimates for all workplace use.
What does OpenAI’s Critical cybersecurity rating mean?
OpenAI’s system card rates both models Critical for cybersecurity risk, the higher of the two thresholds it uses. The supplied source says more permissive cyber access requires extra vetting, but does not specify all access rules or establish that the models have caused harm.
What is Astra for Law, and how does it differ?
OpenAI lists GPT-6.2 Astra for Law as a legal model configuration at Astra’s standard listed prices. The available pages do not explain how it differs from Astra, and Artificial Analysis had not assigned it an Intelligence Index score when its results were accessed.
Source: ThorstenMeyerAI.com
