OpenAI releases GPT-6 Sol and GPT-6 Luna
Teams gain a clear price ladder, but inputs above 272K use higher rates and provider benchmarks do not establish accepted-task cost.
OpenAI GPT-6 Sol and Luna launch ↗L6
Model AI · PROFILE REVIEWED 23 SEPTEMBER 2026
GPT-6 Luna
The low-cost baseline in OpenAI's new family, with the same published context and output limits as Sol. It should be tested first for volume workflows, but HubAI assigns no score until reliability, tool use, latency and accepted-output cost are measured independently.
BEST FORHigh-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context windowNOT IDEAL FORLower token price is not evidence of adequate task quality or lower complete cost
Be the first verified reviewer0 approved reviewsHUBAI DECISION SIGNAL
NEWTest pending
- Price
- $0.10/1M input, $0.01/1M cached input, $0.125/1M cache write and $0.50/1M output at standard API rates; requests above 272K input use higher rates
- Access
- Metered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go users
VERIFIED CHANGE HISTORY
Open Change Radar →What changed for GPT-6 Luna.
EDITOR VERDICT
The low-cost baseline in OpenAI's new family, with the same published context and output limits as Sol. It should be tested first for volume workflows, but HubAI assigns no score until reliability, tool use, latency and accepted-output cost are measured independently.
GPT-6 Luna, explained.
GPT-6 Luna is designed for high-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context window. HubAI treats it as a workflow purchase rather than a novelty: the relevant question is whether it produces an acceptable result repeatedly, with controls your team can understand and a cost that remains predictable after the trial ends.
Best forHigh-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context windowTrialMetered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go users
Where GPT-6 Luna fits—and where it does not.
High-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context window
accepted, tested code changes with lower cycle time and no increase in escaped defects
Lower token price is not evidence of adequate task quality or lower complete cost
Assign the same bounded repository change with tests and security checks.
PRODUCT PROOF
PROFILE UPDATED · 23 SEPTEMBER 2026See the workflow before you buy.
01
Bring one representative input
02
Run the same job three times
03
Measure accepted output and cost
Verify the data-use, retention, model-improvement, regional-processing and administrator terms for the exact OpenAI API, ChatGPT Work, Codex or GitHub Copilot route. OpenAI lists EU data residency only for the standard API tier; product surfaces and third-party routes have separate controls.
gpt-6-luna through the OpenAI API, with gradual rollout in ChatGPT Work, Codex and eligible GitHub Copilot plans. The API page lists a 1,050,000-token context window, 128K maximum output and reasoning from none through max. Inputs above 272K use higher full-request rates.
OpenAI Responses API · ChatGPT Work · Codex · GitHub Copilot · Text and image input · Tool-enabled agent workflows
Vendor facts and HubAI opinion are separated. Sponsored placement cannot buy a higher score.
HUBAI PROCUREMENT PASSPORT
A buying record your team can challenge.
One compact record for price, access, data handling, deployment and the test that must pass before adoption.
BUY FORHigh-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context windowCOMMERCIAL ROUTE$0.10/1M input, $0.01/1M cached input, $0.125/1M cache write and $0.50/1M output at standard API rates; requests above 272K input use higher ratesMetered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go usersTEST PENDING
Verify the data-use, retention, model-improvement, regional-processing and administrator terms for the exact OpenAI API, ChatGPT Work, Codex or GitHub Copilot route. OpenAI lists EU data residency only for the standard API tier; product surfaces and third-party routes have separate controls.
gpt-6-luna through the OpenAI API, with gradual rollout in ChatGPT Work, Codex and eligible GitHub Copilot plans. The API page lists a 1,050,000-token context window, 128K maximum output and reasoning from none through max. Inputs above 272K use higher full-request rates.
OpenAI Responses API · ChatGPT Work · Codex · GitHub Copilot · Text and image input · Tool-enabled agent workflows
Run ten frozen tasks through Luna, then route failures and high-value cases to Sol. Keep tools, permissions and acceptance tests fixed; record accepted results, unsupported output, tool failures, retries, cache behaviour, latency, review minutes and complete billed cost.
No launch-day score.
HubAI will publish a score only after the same representative tasks have been run against current alternatives. Vendor benchmarks are not treated as an editorial rating.
Cost reality
The subscription price is only the starting point. HubAI evaluates usage limits, failed attempts and the work needed to produce one usable result. Pricing changes frequently, so the structure and verification date matter more than a headline price.
Budget for the result, not the plan badge.
- Published structure
- $0.10/1M input, $0.01/1M cached input, $0.125/1M cache write and $0.50/1M output at standard API rates; requests above 272K input use higher rates
- Free or trial route
- Metered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go users
- Primary cost driver
- seat price, premium requests, agent runs and reviewer correction time
- Budget test
- Record the total cost, attempts and human correction needed to produce ten accepted results.
Pricing checked for editorial use on 23 September 2026. Confirm live price, VAT, region, usage rights and cancellation terms with the vendor before purchase.
HUBAI TRUE COST LAB
What does one accepted result really cost?
Model your own usage. HubAI separates subscription cost from retries and human review—without pretending a headline plan price tells the whole story.
DERIVED VALUE · EFFECTIVE COST / ACCEPTED OUTPUT£3.41Human review dominates
PRACTICAL GUIDE
Where GPT-6 Luna earns its place.
Build and test a long-running coding agent
Test this job with your own inputs and judge the finished result—not the first draft.
Run multi-file refactoring with deterministic tool calls
Test this job with your own inputs and judge the finished result—not the first draft.
Evaluate a complex enterprise workflow against cost and latency
Test this job with your own inputs and judge the finished result—not the first draft.
A safer four-step trial
- Start with one narrow job that matches GPT-6 Luna's strongest use case.
- Prepare approved source material, examples and a clear acceptance standard.
- Run three realistic tasks and record attempts, correction time and export quality.
- Keep human approval before publishing, sending or acting on the result.
Start with a better brief.
01
Act as a specialist in model ai. Ask me three questions before creating the first draft. The outcome I need is: [describe outcome].02
Create two approaches for [task]. Use this audience: [audience]. Respect these constraints: [constraints]. Explain the trade-off between the two approaches.03
Review this output against accuracy, tone, privacy, cost and usability. List the changes required before a person should approve it.RISK & GOVERNANCE
What to check before adoption.
Do not place confidential, personal or rights-restricted material into GPT-6 Luna until your organisation has reviewed its current terms, retention settings and account controls. AI output can be plausible and still wrong; regulated, financial, legal and employment decisions require qualified human review.
What HubAI checked—and when.
This is HubAI's editorial record, not the vendor's product changelog or a live uptime claim.
Decision layer expanded
Use-case routes, relevant comparisons and alternatives connected to the product profile.
Commercial profile reviewed
Entry price structure, free or trial route and buyer limitations checked for editorial use.
Editorial profile published
Verdict, best fit, strengths, limitations, prompts and governance questions structured.
Alternatives to GPT-6 Luna
S6GPT-6 SolTest pending
Complex coding, agentic and professional tasks where higher accepted-task quality can justify a premium over a volume model
O5.5Claude Opus 5.5Test pendingLong-running agentic coding and professional knowledge work where migration changes can be tested before production
G4.7Grok 4.7Test pendingLong-context coding, agentic tool use and knowledge work where teams can run a controlled same-task evaluation
GPT-6 Luna against its closest shelf.
Scores are editorial signals, not a universal winner. Choose around the exact job and run the same proof test.
| Product | HubAI score | Best for | Price structure | Access | Decision |
|---|---|---|---|---|---|
| GPT-6 Luna CURRENT | Test pending | High-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context window | $0.10/1M input, $0.01/1M cached input, $0.125/1M cache write and $0.50/1M output at standard API rates; requests above 272K input use higher rates | Metered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go users | Current profile |
| GPT-6 Sol | Test pending | Complex coding, agentic and professional tasks where higher accepted-task quality can justify a premium over a volume model | $2/1M input, $0.20/1M cached input, $2.50/1M cache write and $10/1M output at standard API rates; requests above 272K input use higher rates | Metered API access; gradual availability in ChatGPT Work and Codex for eligible paid, Business, Enterprise and Edu plans, plus selected paid GitHub Copilot plans | Compare → |
| Claude Opus 5.5 | Test pending | Long-running agentic coding and professional knowledge work where migration changes can be tested before production | $4/1M input and $20/1M output at standard API rates; cache reads cost $0.20/1M, five-minute writes $5/1M and one-hour writes $8/1M; Fast mode is $8/$40 | Paid API and cloud-platform access; availability across Claude plans varies by entitlement and route | Compare → |
| Grok 4.7 | Test pending | Long-context coding, agentic tool use and knowledge work where teams can run a controlled same-task evaluation | $2/1M input and $6/1M output tokens on the standard xAI API; the US regional endpoint adds 10%, while the Fast route in Cursor and Grok Build uses higher rates | Grok Build has a free starting route; public API use is metered. GitHub Copilot access requires an eligible paid plan and is rolling out gradually | Compare → |
Questions to answer before adoption.
- 01What repository data is retained or used for training?
- 02Can policies block secrets and unsafe commands?
- 03Which IDE and Git workflows are supported?
- 04How much reviewer time does an accepted change need?
GPT-6 Luna FAQ.
What is GPT-6 Luna best for?
GPT-6 Luna is best suited to high-volume extraction, classification, transformation and first-pass agent work that needs low token cost and a large context window. HubAI's current verdict is: The low-cost baseline in OpenAI's new family, with the same published context and output limits as Sol. It should be tested first for volume workflows, but HubAI assigns no score until reliability, tool use, latency and accepted-output cost are measured independently.
Does GPT-6 Luna have a free plan or trial?
Metered API access; gradual rollout through ChatGPT Work, Codex and eligible GitHub Copilot plans, with Luna desktop access announced for Free and Go users. Check the official product page before buying because allowances, regions and eligibility can change.
How much does GPT-6 Luna cost?
$0.10/1M input, $0.01/1M cached input, $0.125/1M cache write and $0.50/1M output at standard API rates; requests above 272K input use higher rates. The useful comparison is total cost per accepted result, including retries, limits and human correction—not the advertised entry price alone.
What should I compare with GPT-6 Luna?
Compare products in the same Code AI shelf against one identical task, acceptance rule and cost window. Focus on accepted, tested code changes with lower cycle time and no increase in escaped defects.
VERIFIED COMMUNITY SIGNAL
Did you like GPT-6 Luna?
Rate this product and share what really happened. Every review is checked before it appears publicly.
HubAI asks
Did this product earn its cost?
If you used GPT-6 Luna, share its quality, pricing, trial experience and real result. Useful specifics beat empty praise.
Sign in to leave a rating and review.
Sign in to reviewNo approved user reviews yet. Be the first to leave a useful one.