How many GPT-6 Luna messages do you get in Codex and ChatGPT Work?
OpenAI estimates 350 to 3,000 GPT-6 Luna local messages every five hours on Plus, 1,750 to 14,000 on Pro 5x and 7,000 to 56,000 on Pro 20x, shared between Codex and ChatGPT Work.
Checked against the lab's own pages on September 26, 2026. Limits change often, so the date matters.
On ChatGPT Plus, OpenAI's pricing page estimates 350 to 3,000 GPT-6 Luna local messages every five hours in Codex. Pro 5x gets 1,750 to 14,000 and Pro 20x gets 7,000 to 56,000. That allowance is shared with ChatGPT Work, which uses the same pricing, credits and usage limits as Codex.
Those are the biggest numbers in OpenAI's table by a long way. They're still estimates, though, and a long agent session will use far more per message than a small script.
- Plus
- 350-3,000 local messages per five-hour period
- Pro 5x
- 1,750-14,000
- Pro 20x
- 7,000-56,000
- Standard Business
- 350-3,000
- Credits
- 2.5 credits per 1M input tokens, 12.5 credits per 1M output tokens
- Fast mode
- 2.5x the Standard credit rate
How far does Luna go next to Sol and Astra?
About twenty times further than Sol, going by the same table. GPT-6 Sol tops out at 150 messages on Plus, and Luna's range tops out at 3,000. The credit rates line up with that: Luna costs 2.5 credits per million input tokens, and Sol costs 50.
Astra sits at the other end, at 5 to 45 messages on Plus. And GPT-5.6 Luna is listed at 250 to 2,000 on Plus, so GPT-6 Luna's range is higher at both ends.
These estimates are not fixed message limits; check your usage dashboard for current limits and reset times.
Does the free plan get Luna?
Yes, as it rolls out. Free and Go both list GPT-6 Luna at Standard speed in the desktop app, subject to rollout. OpenAI doesn't print a message estimate for either plan, so there's no number to give you for them.
Business ($100) uses the Pro 5x estimates. Enterprise and Edu with flexible pricing have no fixed rate limits at all; usage scales with credits.
What uses it up faster
Fast mode uses 2.5x the Standard credit rate for Luna where it's available, so included limits go faster. Image generation uses them about 3 to 5 times faster on average. Local messages and cloud chats share one allowance, and weekly limits may also apply on top of the five-hour window.
When you run out
If you hit the limit mid-turn, the agent can finish that turn, subject to fair use limits. Plus and Pro users can then buy credits without upgrading, and anyone can keep going on an API key at API rates. The /status command shows what's left during a Codex CLI session. For the other models, see GPT-6 Sol's limits and the API rate limits for Sol and Luna.
These numbers were read on September 26, 2026.