Official rollout · Facts verified 01 Oct 2026
GPT-6 Sol: Availability, Limits and GPT-6 Luna
Looking for GPT-6.1 Sol? On September 29, 2026, OpenAI introduced GPT-6.1 Sol (gpt-6.1-sol) as the newer Sol model, featuring near-Astra performance and bundled default status in Codex CLI 0.159.1. Original GPT-6 Sol remains active and available. Read our dedicated GPT-6.1 Sol usage, limits & configuration guide.
GPT-6 Sol is real and is rolling out in Codex and ChatGPT Work. OpenAI announced Sol and GPT-6 Luna on 22 Sep 2026. Choose gpt-6-sol for complex coding and agentic work, or gpt-6-luna for focused, repeatable tasks. This is OpenAI's recommendation; it is not a conclusion from our own benchmarks.
Both models are rolling out to Plus, Pro, Business, Enterprise and Edu. Free and Go users can access Luna in the desktop app at Standard speed, subject to rollout. Enterprise administrators must enable access. In ChatGPT, Sol and Luna are available in Work and Codex, not Chat.
Official Codex estimates and credit rates
| Model | Plus | Pro $100 | Previous Pro 20x estimates | Business Standard |
|---|---|---|---|---|
GPT-6 Solgpt-6-sol | 15–150 | 70–700 | 300–3,000 | 15–150 |
GPT-6 Lunagpt-6-luna | 350–3,000 | 1,750–14,000 | 7,000–56,000 | 350–3,000 |
GPT-6 Astragpt-6-astra | 5–45 | 25–225 | 100–900 | 5–45 |
OpenAI estimates of local messages per five-hour period, not fixed caps. Actual usage varies; weekly limits may also apply. Pro $100, $200, $500 and Business Premium have no five-hour usage limit. The Pro columns preserve earlier published estimates, not current caps. Pro plans still have included usage allowances; check Settings > Usage for remaining allowance and reset times. Oldest verification in these tables: 2026-09-27. OpenAI usage guidance.
Pro $200 is available for new subscriptions with a lower allowance for non-grandfathered accounts; OpenAI has not published its replacement multiplier. Eligible grandfathered subscribers keep the previous allowance through October 29, 2026 while subscribed, then move to the lower allowance. Earlier published Pro 5x and Pro 20x message estimates are historical evidence, not current five-hour caps or proof of current Pro $200 entitlement. Pro $100, $200, and $500 currently have no five-hour usage limit in Work and Codex, but still have included usage allowances. Pro $500 has the highest included Pro usage and Astra Ultrafast, with no published multiplier or model-specific message estimate. OpenAI Pro tiers.
Free Luna access: desktop only, Standard speed; subject to rollout. No numeric estimate is published.
Go Luna access: desktop only, Standard speed; subject to rollout. No numeric estimate is published.
Astra access: Plus: Limited Astra usage in Work and Codex; Business Standard: Limited Astra usage in Work and Codex; Pro $100: Full existing Work and Codex allowance; Pro $200: Full existing Work and Codex allowance; Pro $500: Full existing Work and Codex allowance; Business Premium: Full existing Work and Codex allowance; Enterprise: Workspace agreement and enablement apply; Edu: Workspace agreement and enablement apply.
| Standard speed, per 1M tokens | Input | Cached input | Output |
|---|---|---|---|
| GPT-6 Sol | 50 credits | 5 credits | 250 credits |
| GPT-6 Luna | 2.5 credits | 0.25 credits | 12.5 credits |
| GPT-6 Astra | 250 credits | 25 credits | 1,250 credits |
ChatGPT Work and Codex credit rates, not API dollars or a formula for included allowance. How credits work.
These models draw from the same Work and Codex allowance. Switching models does not restore exhausted usage. See plan limits and reset mechanics.
API pricing and context are separate
At Standard speed, Sol's published API rates per million tokens are $2 input, $0.2 cached input, $2.5 cache writes and $10 output. Luna's rates are $0.1, $0.01, $0.125 and $0.5, respectively. These are USD API charges, not ChatGPT credit rates.
Both API model pages publish a 1,050,000-token context window and 128,000 maximum output tokens. Neither number is your subscription quota or a guaranteed context size in every Codex client. Inputs over 272K tokens raise the full request's input and cache rates to 2x and output to 1.5x; regional processing, tools and speed settings can add costs. Check the Sol API page and Luna API page.
Select Sol or Luna in Codex
Open the model picker or use /model. OpenAI's CLI 0.156.1 release notes added picker support on 23 Sep 2026. Update your client if the models are missing, then check your account and workspace access.
codex --model gpt-6-sol
codex --model gpt-6-lunaFor defaults and saved configuration, read recommended model settings. Defaults vary by client and account; OpenAI does not document a universal prompt-by-prompt model router.
Earlier models, research and Reserve
GPT-5.6 Sol, Terra and Luna remain available during rollout. Our controlled quota measurements tested GPT-5.6 models and Astra under their recorded conditions. They do not measure GPT-6 Sol or GPT-6 Luna and are not official limits.
The Luna limits page distinguishes generations. OpenAI's Luna Reserve guidance still identifies GPT-5.6 Luna; we have not found an official Reserve migration announcement. See the changes ledger for the rollout and migration history.
Earlier versions of this page described Sol as unconfirmed before launch. That statement became obsolete after the September 22 announcement and was corrected on September 27.
Common questions
Is GPT-6 Sol real and released?
Yes. OpenAI announced the GPT-6 Sol and GPT-6 Luna rollout on 2026-09-22. The model IDs are gpt-6-sol and gpt-6-luna. Availability still depends on rollout, client and workspace settings.
Is GPT-6 Sol available in ChatGPT?
It is available in ChatGPT Work and Codex, not ordinary Chat. Sol and Luna are rolling out to Plus, Pro, Business, Enterprise and Edu. Free and Go can access Luna in the desktop app, subject to rollout. Enterprise administrators must enable the new models.
How many Codex messages can I send with GPT-6 Sol?
OpenAI estimates 15–150 local messages per five-hour period on Plus and Business Standard, 70–700 in the earlier Pro 5x column, and 300–3,000 in the published legacy Pro 20x column. This does not establish every current Pro $200 account's allowance. The replacement Pro $200 allowance and Pro $500 message estimates are unpublished. These are estimates, not fixed quotas. Pro $100, $200, and $500 currently have no five-hour usage limit in Work and Codex, but still have included usage allowances; check Settings > Usage for remaining allowance and reset times. Weekly limits may also apply.
Does GPT-6 Sol replace GPT-5.6 Sol?
GPT-6 Sol is the newer recommendation for complex coding. OpenAI says GPT-5.6 Sol, Terra and Luna remain available during rollout. That does not establish a GPT-5.6 Sol retirement date. Current migration guidance points paid Codex users on retiring GPT-5.5 to GPT-6 Sol when available.
Does GPT-6 Luna replace Luna Reserve?
OpenAI's Reserve article still names GPT-5.6 Luna. The GPT-6 Luna rollout does not establish that Reserve has migrated. Reserve remains a separate allowance for selected personal Plus and Pro accounts.
Sources: OpenAI changelog, model documentation, pricing documentation.