• About
  • Advertise
  • Privacy Policy
  • Contact
  • Editorial Policy
  • Corrections Policy
TrustLoud
  • Home
  • AI News
  • AI
    • AI Agents & Automation
    • AI for WordPress & Web
  • Tech
    • Gadgets
    • Smartphones
    • Laptops & PCs
    • Software & Apps
    • Cybersecurity
  • Reviews
  • Comparisons
  • How-To
No Result
View All Result
  • Home
  • AI News
  • AI
    • AI Agents & Automation
    • AI for WordPress & Web
  • Tech
    • Gadgets
    • Smartphones
    • Laptops & PCs
    • Software & Apps
    • Cybersecurity
  • Reviews
  • Comparisons
  • How-To
No Result
View All Result
TrustLoud
No Result
View All Result
Home AI News: What Changed

GPT-6.2 Astra and Sol: Release Date, Pricing, Benchmarks and Who Gets It

Veer by Veer
November 18, 2026
144 6
TrustLoud title graphic reading GPT-6.2 Astra and Sol, Price, Benchmarks, Access, with the subtitle What OpenAI launched on November 17, 2026
Share on FacebookShare on Twitter

TL;DR: OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026. Astra is the new flagship “for your hardest projects,” and Sol is the cheaper everyday model. In the API, GPT-6.2 Astra costs $8 per million input tokens and $40 per million output tokens, 20% less than GPT-6 Astra’s $10 / $50. GPT-6.2 Sol stays at $2 / $10, the same as GPT-6.1 Sol. In ChatGPT, both models are rolling out to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and Codex only. OpenAI’s Help Center says the rollout “does not add either model to regular Chat,” and free and Go accounts aren’t included. On Artificial Analysis’ independent Intelligence Index, GPT-6.2 Astra scores 62 (up from GPT-6 Astra’s 57) at about $4 per task. Anthropic’s Claude Fable 5.2 still scores higher at 65, but costs about $26.54 per task.

Last updated: November 18, 2026. By Veer. Facts checked on November 18, 2026 against OpenAI’s GPT-6.2 announcement, the GPT-6.2 Astra and GPT-6.2 Sol API model pages, the API changelog, the ChatGPT Work and Codex pricing doc, the OpenAI Help Center and Artificial Analysis. This is a researched news explainer built from public documents. OpenAI’s benchmark numbers are the company’s own claims.

GPT-6.2 arrived just over two months after GPT-6 Astra (September 3) and seven weeks after GPT-6.1 Sol (September 29). It also landed about a week after Anthropic’s Claude Fable 5.2, which 9to5Google frames as the model OpenAI is chasing. Autocomplete shows what people want to know: “gpt 6.2 release date,” “gpt 6.2 astra pricing,” “gpt 6.2 benchmarks” and “gpt 6.2 astra vs fable 5.2.” This post answers each one.

When was GPT-6.2 released?

GPT-6.2 Astra and GPT-6.2 Sol were released on Tuesday, November 17, 2026, in ChatGPT, Codex and the API on the same day. OpenAI’s API changelog lists gpt-6.2-astra and gpt-6.2-sol under November 17. Both support the Responses API and Chat Completions, and tool calling needs the Responses API.

The ChatGPT rollout is gradual. The Help Center says to update the ChatGPT desktop app (Menu > Check for Updates) and the Codex CLI, then restart. If the models still don’t show up, “they may not be available for your account yet.”

Model Released API model ID Role
GPT-6 Astra September 3, 2026 gpt-6-astra Previous flagship
GPT-6.1 Sol September 29, 2026 gpt-6.1-sol Previous everyday model
GPT-6.2 Astra November 17, 2026 gpt-6.2-astra New flagship, “for your hardest projects”
GPT-6.2 Sol November 17, 2026 gpt-6.2-sol New everyday model

Sources: OpenAI API changelog, GPT-6.2 announcement, 9to5Mac.

How much does GPT-6.2 cost?

In the API, GPT-6.2 Astra is $8 input / $0.40 cached input / $40 output per million tokens, and GPT-6.2 Sol is $2 / $0.10 / $10. OpenAI says Astra’s input and output prices are 20% lower than GPT-6 Astra’s, with “a deeper 60% discount on cached input tokens.” Sol “carries forward GPT-6.1 Sol’s rates.”

Per 1M tokens (Standard) GPT-6.2 Astra GPT-6 Astra GPT-6.2 Sol GPT-6.1 Sol
Input $8.00 $10.00 $2.00 $2.00
Cached input $0.40 $1.00 $0.10 $0.10
Output $40.00 $50.00 $10.00 $10.00
Cache write (30-minute) $10.00 n/a $2.50 n/a
Context window / max output 1.05M / 128K 1.05M / 128K 1.05M / 128K 1.05M / 128K
Knowledge cutoff May 18, 2026 April 30, 2026 June 1, 2026 April 30, 2026

Sources: GPT-6.2 Astra model page, GPT-6.2 Sol model page, OpenAI API pricing; GPT-6 Astra and GPT-6.1 Sol figures from our GPT-6.1 Sol vs GPT-6 Astra comparison. All prices USD.

A few billing details from the model pages are worth knowing before you switch:

  • Long prompts cost more. Above 272K input tokens, input and cache rates double and output costs 1.5x for the whole request.
  • Batch is half price; Fast is double. Both apply to the GPT-6.2 models.
  • Ultrafast is cheaper for now. API Ultrafast costs 3x the Standard rate through February 28, 2027, then 6x from March 1, 2027 (Pacific time). OpenAI says Ultrafast generates tokens at up to 4.5 times the standard rate, per 9to5Mac.
  • 12-hour caching costs extra. A 12-hour cache write is $20 per million tokens on Astra and $5 on Sol.

For a rough sense of scale, a job that reads 1 million tokens and writes 200,000 costs about $16 on GPT-6.2 Astra, against $20 on GPT-6 Astra and $4 on GPT-6.2 Sol, at standard rates without caching.

Who can use GPT-6.2 in ChatGPT?

Paid ChatGPT users on Plus, Pro, Business, Enterprise and Edu get GPT-6.2 Astra and Sol in ChatGPT Work and Codex, not in regular Chat. OpenAI’s launch post names Plus, Pro, Enterprise and Edu. Its pricing doc and Help Center also list Business, and the Help Center adds ChatGPT for Healthcare and FedRAMP workspaces “as access becomes available.”

  • Not in Chat. “This rollout does not add either model to regular Chat,” the Help Center says. To use them, open Work and pick the model, or use the Codex app or CLI (codex -m gpt-6.2-astra).
  • Free and Go: not included in the launch.
  • Enterprise and Edu: both models are off by default until an admin turns them on.
  • New Work chats default to GPT-6.2 Sol at the Light setting on supported consumer plans, according to OpenAI’s models guide.
  • Ultrafast: the $500/month Pro plan gets it on both models. The $200/month Pro plan gets Ultrafast for the first time, starting with GPT-6.2 Sol.

How many GPT-6.2 messages do you get on Plus?

OpenAI estimates 6 to 50 GPT-6.2 Astra messages and 17 to 180 GPT-6.2 Sol messages per five hours on Plus. That’s a bit more than the old models (5 to 45 for GPT-6 Astra, 15 to 160 for GPT-6.1 Sol), which fits OpenAI’s claim that the new models use fewer output tokens. These are estimates for local messages, not fixed limits, and weekly limits may also apply. Pro plans currently have no five-hour limit.

ChatGPT Work and Codex (Plus) GPT-6.2 Astra GPT-6 Astra GPT-6.2 Sol GPT-6.1 Sol
Messages per 5 hours (estimate) 6–50 5–45 17–180 15–160
Credits per 1M input / output tokens 200 / 1,000 250 / 1,250 50 / 250 50 / 250
Ultrafast drain on included usage 6x 8x 6x 8x

Source: ChatGPT Work and Codex pricing, checked November 18, 2026.

If you’re deciding which plan to buy for this, our ChatGPT Pro $100 vs $200 vs $500 guide covers the Pro tiers.

GPT-6.2 benchmarks: how good is it?

On Artificial Analysis’ independent Intelligence Index, GPT-6.2 Astra scores 62 and GPT-6.2 Sol scores 57, up from 57 for GPT-6 Astra and 55 for GPT-6.1 Sol. The bigger change is cost and speed. GPT-6.2 Astra costs about $4.00 per index task against $6.24 for GPT-6 Astra, and finishes in about 11.5 minutes against 19.9.

Artificial Analysis (max effort) Intelligence Index Cost per task Minutes per task
Claude Fable 5.2 65 $26.54 41.1
GPT-6.2 Astra 62 $4.00 11.5
GPT-6.2 Sol 57 $1.01 7.9
GPT-6 Astra 57 $6.24 19.9
GPT-6.1 Sol 55 $1.27 23.3

Sources: Artificial Analysis model pages for GPT-6.2 Astra, GPT-6.2 Sol, GPT-6 Astra, GPT-6.1 Sol and Claude Fable 5.2, Intelligence Index v5.0.1, checked November 18, 2026. Fable 5.2’s listing is its “Max, Default Fallback” configuration.

OpenAI’s own launch numbers go further. They’re vendor results, so treat them as claims:

  • Agents’ Last Exam (multistep professional work): 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 Claude Fable 5.2 “at just 5% of the cost.” OpenAI notes Fable 5.2’s leaderboard results exclude nine tasks that used a fallback model.
  • Terminal-Bench-Science 0.1: Astra scores 9 percentage points higher than GPT-6 Astra. OpenAI calls it “the best model in the world for scientists.”
  • Cognition’s FrontierCode 2.0 (code changes good enough to merge): GPT-6.2 Sol comes within 0.5 points of GPT-6.2 Astra “at about one-fifth the cost.”
  • GDPval v2 (long knowledge-work projects): Astra is up 6.2 points on GPT-6 Astra while using 22% fewer output tokens; Sol is up 5.3 points on GPT-6.1 Sol.
  • AA-3D (human votes on browser 3D experiences): Astra ranks first at 1,318 Elo, ahead of Claude Fable 5.2 (1,271) and GPT-6.2 Sol (1,259).

All five claims come from OpenAI’s GPT-6.2 announcement.

GPT-6.2 Astra vs Claude Fable 5.2

Fable 5.2 still leads on Artificial Analysis’ overall index (65 vs 62), but GPT-6.2 Astra gets close at less than a sixth of the cost per task. Both list at the same token price, $8 input and $40 output per million, according to Artificial Analysis. The gap comes from token use. Fable 5.2 generated 275 million tokens across the index against 65.8 million for GPT-6.2 Astra, and took 41.1 minutes per task against 11.5.

So the choice depends on the job. If you need the highest score and cost isn’t the constraint, Fable 5.2 is still ahead on the broad independent benchmark. If you run a lot of agent or coding work, GPT-6.2 Astra (or Sol) does most of it for far less. On specific professional benchmarks, OpenAI claims Astra beats Fable, including Agents’ Last Exam and Harvey’s legal benchmark, where it says Astra more than doubles Fable 5.2’s pass rate at about one-sixth the cost. Those aren’t independent results yet. If you’re on Anthropic’s side of the fence, our Sonnet 5.5 vs Opus 5.5 guide and Claude Max price increase explainer cover the current Claude lineup and plans.

What’s new in GPT-6.2 besides benchmarks?

OpenAI’s headline changes are better documents and slides, cleaner code, sharper computer use, and stronger cybersecurity safeguards. Here’s the short list from the announcement:

  • Slides and documents. Astra can follow a slide template from an original file “or even a screenshot,” matching layout, colors and font size. OpenAI says its writing has “fewer generic takeaways and less filler.”
  • Coding. OpenAI says Astra is better at implementing complex changes “in existing codebases,” and that early testers reported “cleaner code style” with “fewer rounds of revision.”
  • Computer use and 3D. More precise control in CAD and simulation tools, and better web and 3D design.
  • Cybersecurity limits. Both models meet OpenAI’s “Critical” cybersecurity threshold, so they launch “with tighter safeguards for higher-risk cybersecurity assistance.” OpenAI admits this will “restrict some legitimate vulnerability research and security” work. GPT-6.2 isn’t on OpenAI’s Daybreak program for defenders yet.
  • AI self-improvement rating. GPT-6.2 Astra is OpenAI’s “first externally released model to meet the High threshold for AI Self-improvement.”
  • A known quirk. In two alignment evaluations, both models “were more likely to omit mentioning obstacles that blocked their work, such as a missing file.” OpenAI says it added Codex instructions that “substantially reduce” this and is still investigating.

For developers: the changelog says GPT-6.2 “changes how the Responses API preserves reasoning when you edit a conversation’s prefix and how it handles the top-level instructions field.” OpenAI recommends keeping prompt prefixes unchanged and appending new instructions to improve cache reuse. The API also rejects requests that set reasoning effort to none.

Should you switch to GPT-6.2?

Yes, for most paid ChatGPT and API users there’s little reason to stay on the older models. GPT-6.2 Sol costs the same as GPT-6.1 Sol and scores higher. GPT-6.2 Astra is cheaper per token than GPT-6 Astra and, on Artificial Analysis’ figures, faster and cheaper per task too.

  • Everyday coding, writing and agents: use GPT-6.2 Sol. It’s the new default in Work and the best value in OpenAI’s lineup.
  • Hardest projects, science and polished deliverables: use GPT-6.2 Astra.
  • API users: read OpenAI’s GPT-6.2 migration guide first because of the Responses API changes, and re-check prompts that depend on prefix edits or the instructions field.
  • Security researchers: expect more refusals on higher-risk requests until GPT-6.2 comes to Daybreak.
  • Free and Go users: nothing changes for now.

For a deeper look at how Sol and Astra trade off, see our GPT-6.1 Sol vs GPT-6 Astra comparison; the same logic carries over to the 6.2 models. Coders comparing plans can also check Claude Code vs Codex.

FAQ

When did GPT-6.2 come out?

OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026, in ChatGPT Work, Codex and the API on the same day. The ChatGPT rollout is gradual, so some accounts may see the models later.

How much does GPT-6.2 Astra cost?

In the API, GPT-6.2 Astra costs $8 per million input tokens, $0.40 per million cached input tokens and $40 per million output tokens. That’s 20% less than GPT-6 Astra on input and output. GPT-6.2 Sol costs $2, $0.10 and $10.

Is GPT-6.2 available for free ChatGPT users?

No. GPT-6.2 Astra and Sol are rolling out to Plus, Pro, Business, Enterprise and Edu in ChatGPT Work and Codex. OpenAI’s Help Center says the rollout does not add either model to regular Chat.

What is the difference between GPT-6.2 Astra and GPT-6.2 Sol?

Astra is OpenAI’s most capable model for the hardest projects. Sol is the cheaper everyday model at a quarter of Astra’s token price. Artificial Analysis scores Astra 62 and Sol 57, and OpenAI says Sol comes within 0.5 points of Astra on FrontierCode 2.0 at about one-fifth the cost.

Is GPT-6.2 Astra better than Claude Fable 5.2?

It depends on the measure. Claude Fable 5.2 scores 65 to GPT-6.2 Astra’s 62 on Artificial Analysis’ Intelligence Index, but costs about $26.54 per task against $4.00. OpenAI’s own benchmarks show Astra ahead of Fable 5.2 on several professional tasks.

How many GPT-6.2 messages do I get on ChatGPT Plus?

OpenAI estimates 6 to 50 GPT-6.2 Astra messages and 17 to 180 GPT-6.2 Sol messages per five hours on Plus in Work and Codex. These are estimates, weekly limits may also apply, and Pro plans have no five-hour limit.

Sources

  1. Introducing GPT-6.2 Astra and Sol (OpenAI, November 17, 2026): benchmarks, safety, pricing and availability.
  2. GPT-6.2 Astra and GPT-6.2 Sol API model pages: prices, context, knowledge cutoffs, billing rules.
  3. OpenAI API changelog: November 17 release and Responses API changes.
  4. ChatGPT Work and Codex pricing and models guide: plans, message estimates, credits, Ultrafast.
  5. ChatGPT Work and Codex (OpenAI Help Center): rollout and “not in regular Chat” note.
  6. Artificial Analysis: Intelligence Index scores, cost and time per task (checked November 18, 2026).
  7. 9to5Mac and 9to5Google (November 17, 2026): launch coverage and Pro plan details.

Tags: AI ModelsChatGPTCodexGPT-6OpenAI
Veer

Veer

Veer is the editor of TrustLoud. Veer researches and writes explainers, comparisons and how-to guides on AI tools and consumer tech, checking prices and specs against official sources before publishing. Every post lists its sources and a last-updated date. Spotted an error? Use the Contact page and we'll correct it.

Recent Posts

  • How to Screenshot on a Chromebook: Every Shortcut and Where Files Go
  • How to Screenshot on Windows 11 and 10: Every Shortcut and Where Files Go
  • GPT-6.2 Astra and Sol: Release Date, Pricing, Benchmarks and Who Gets It
  • How to Screenshot on Mac: Every Shortcut, Plus Where Files Go (2026)
  • How to Block Websites on iPhone: Screen Time, Kids and DNS (2026)

Recent Comments

  1. How to Screenshot on a Chromebook (Every Shortcut) on How to Screenshot on Mac: Every Shortcut, Plus Where Files Go (2026)
  2. How to Screenshot on a Chromebook: Every Shortcut and Where Files Go - TrustLoud on How to Screenshot on Windows 11 and 10: Every Shortcut and Where Files Go
  3. How to Screenshot on Windows 11 and 10 (Every Shortcut) on How to Screenshot on Mac: Every Shortcut, Plus Where Files Go (2026)
  4. GPT-6.1 Sol vs GPT-6 Astra: Price and Which to Use on Sonnet 5.5 vs Opus 5.5: Pricing, Speed and Which to Pick
  5. GPT-6.2 Astra and Sol: Pricing, Benchmarks, Access on GPT-6.1 Sol vs GPT-6 Astra: Price, Performance and Which to Use (2026)

Archives

  • November 2026
  • October 2026

Categories

  • AI Agents & Automation
  • AI for WordPress & Web
  • AI News: What Changed
  • Comparisons
  • Cybersecurity
  • Gadgets
  • How-To Guides
  • Laptops & PCs
  • Smartphones
  • Software & Apps

Categories

  • AI Agents & Automation
  • AI for WordPress & Web
  • AI News: What Changed
  • Comparisons
  • Cybersecurity
  • Gadgets
  • How-To Guides
  • Laptops & PCs
  • Smartphones
  • Software & Apps

Tags

  • Screenshots

  • Parental Controls

  • Screen Time

  • United Airlines

  • Safari

  • Factory Reset

  • Flight Refunds

Recent News

TrustLoud title graphic reading How to Screenshot on a Chromebook, with the subtitle Every shortcut, screen recording and where files save

How to Screenshot on a Chromebook: Every Shortcut and Where Files Go

November 18, 2026
TrustLoud title graphic reading How to Screenshot on Windows, with the subtitle Every shortcut, Snipping Tool and where files save

How to Screenshot on Windows 11 and 10: Every Shortcut and Where Files Go

November 18, 2026
  • About
  • Advertise
  • Privacy Policy
  • Contact
  • Editorial Policy
  • Corrections Policy

© 2026 TrustLoud. All rights reserved.

No Result
View All Result
  • Home
  • AI News
  • AI
    • AI Agents & Automation
    • AI for WordPress & Web
  • Tech
    • Gadgets
    • Smartphones
    • Laptops & PCs
    • Software & Apps
    • Cybersecurity
  • Reviews
  • Comparisons
  • How-To

© 2026 TrustLoud. All rights reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In

Add New Playlist