Free
$0/mo
- Annual billing
- n/a
- Seats
- individual
- Codex usage
- Limited
Deep guide
OpenAI Codex: plans, limits & models
Every fact below is from an official OpenAI page and is linked and dated. Anything OpenAI does not publish is marked Not published by OpenAI.
GPT-6.1 Sol
GPT-6.1 Sol ("Sol 6.1"), OpenAI's recommended Codex model for complex work.
Latest genGPT-6.1 Sol (API/config id: gpt-6.1-sol)
'near-Astra performance for complex work at a lower cost than Astra'; recommended for complex coding/agentic work
Leaderboard results for GPT-6.1 Sol →- In Codex since
- Who gets it
- Plus, Pro, Business, Enterprise, Edu in Codex (desktop app, CLI) and ChatGPT Work (web, mobile); Enterprise/Edu off by default until admin enables; not Free/Go at launch; not in ChatGPT Chat
- Reasoning efforts
- Light to Ultra; Max and Ultra depend on settings
- Ultrafast
- 2026-10-08: GPT-6.1 Sol Ultrafast on Pro $500 and eligible Enterprise/Edu; inference residency US and Europe (EEA+Switzerland)
Sources developers.openai.com/codex/models · developers.openai.com/codex/changelog Checked
Plans
Prices are USD per month as listed by OpenAI. Business is per seat.
Go
$8/mo
- Annual billing
- Not published by OpenAI
- Seats
- individual
- Codex usage
- Limited
Plus
$20/mo
- Annual billing
- Not published by OpenAI
- Seats
- individual
- Codex usage
- Local messages per 5h (estimates): GPT-6 Astra 5-45; GPT-6.1 Sol 15-160; GPT-6 Sol 15-150; GPT-6 Luna 350-3,000; weekly limits may also apply
Pro 100
$100/mo
- Annual billing
- Pro is billed monthly; annual billing not available
- Seats
- individual
- Codex usage
- Pro plans currently have no five-hour limit; weekly limits may apply
Pro 200
$200/mo
- Annual billing
- monthly only
- Seats
- individual
- Codex usage
- More usage than Pro 100; new non-grandfathered subscriptions get a lower allowance than before; grandfathered allowance through 2026-10-29
Pro 500
$500/mo
- Annual billing
- monthly only
- Seats
- individual
- Codex usage
- Highest included usage of the Pro plans; no 5h limit
Business Standard seat
$25/mo
- Annual billing
- $20/user/month billed annually
- Seats
- 2 seat minimum; up to 200 paid seats per subscription (from 2026-08-24); mix with Premium
- Codex usage
- Same local-message estimates as Plus
Business Premium seat
$125/mo
- Annual billing
- $100/user/month billed annually
- Seats
- 2 seat minimum (mixed)
- Codex usage
- 5x Standard usage, no 5-hour limit
Business Codex seat
usage-based, no fixed per-user cost; requires workspace credits; no minimum
Enterprise / Edu
contact sales
- Codex usage
- With flexible pricing: no fixed rate limits, usage scales with credits; without: same per-seat limits as Plus for most features
API key
pay per token at API pricing
All plan details in one table11
| Plan | Price / month | Annual billing | Seats | Codex models | Default model | Codex usage | Note | GPT-6 Astra | Ultrafast | Pro messages | Regions |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Free | $0 | n/a | individual | GPT-6 Luna at Standard speed in the desktop app, subject to rollout | GPT-6 Luna (only model listed for Free) | Limited | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI |
| Go | $8 | Not published by OpenAI | individual | GPT-6 Luna at Standard speed in the desktop app, subject to rollout | GPT-6 Luna (only model listed for Go) | Limited | This plan may include ads | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI |
| Plus | $20 | Not published by OpenAI | individual | GPT-6.1 Sol and GPT-6 Luna (plus GPT-6 Sol, GPT-6 Astra limited, GPT-5.6 family during rollout) | GPT-6.1 Sol when available (recommended) | Local messages per 5h (estimates): GPT-6 Astra 5-45; GPT-6.1 Sol 15-160; GPT-6 Sol 15-150; GPT-6 Luna 350-3,000; weekly limits may also apply | Not published by OpenAI | limited Astra usage, credits for more | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI |
| Pro 100 | $100 | Pro is billed monthly; annual billing not available | individual | Not published by OpenAI | Not published by OpenAI | Pro plans currently have no five-hour limit; weekly limits may apply | Not published by OpenAI | full existing allowance for Astra | Not included | Not published by OpenAI | Not published by OpenAI |
| Pro 200 | $200 | monthly only | individual | Not published by OpenAI | Not published by OpenAI | More usage than Pro 100; new non-grandfathered subscriptions get a lower allowance than before; grandfathered allowance through 2026-10-29 | Not published by OpenAI | Not published by OpenAI | Not included | Not published by OpenAI | Not published by OpenAI |
| Pro 500 | $500 | monthly only | individual | Not published by OpenAI | Not published by OpenAI | Highest included usage of the Pro plans; no 5h limit | Not published by OpenAI | Not published by OpenAI | Included (GPT-6 Astra and GPT-6.1 Sol; 8x included usage, 6x credit rate) | Not published by OpenAI | Not published by OpenAI |
| Business Standard seat | $25 | $20/user/month billed annually | 2 seat minimum; up to 200 paid seats per subscription (from 2026-08-24); mix with Premium | Not published by OpenAI | Not published by OpenAI | Same local-message estimates as Plus | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | 15 Pro messages per month shared across GPT-6 Pro and GPT-5.6 Sol Pro | pricing may vary by country and currency |
| Business Premium seat | $125 | $100/user/month billed annually | 2 seat minimum (mixed) | Not published by OpenAI | Not published by OpenAI | 5x Standard usage, no 5-hour limit | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | 50 Pro messages per week shared across GPT-6 Pro and GPT-5.6 Sol Pro | Not published by OpenAI |
| Business Codex seat | usage-based, no fixed per-user cost; requires workspace credits; no minimum | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Introduced 2026-04-02 | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI |
| Enterprise / Edu | contact sales | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | With flexible pricing: no fixed rate limits, usage scales with credits; without: same per-seat limits as Plus for most features | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Data residency in US, EU, UK, JP, CA, KR, SG, IN, AU, UAE (Enterprise only) |
| API key | pay per token at API pricing | Not published by OpenAI | Not published by OpenAI | follows API models available to your key | Not published by OpenAI | Not published by OpenAI | CLI, SDK, IDE extension only; no cloud features (GitHub code review, Slack) | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI | Not published by OpenAI |
Sources developers.openai.com/codex/pricing · openai.com/chatgpt/pricing · developers.openai.com/codex/models · help.openai.com/en/articles/20001275-chatgpt-work-and-codex · help.openai.com/en/articles/9793128-what-is-chatgpt-pro · help.openai.com/en/articles/8792536-managing-billing-and-seats-in-chatgpt-business · help.openai.com/en/articles/12003714-chatgpt-business-models-and-limits · help.openai.com/en/articles/8792828-chatgpt-business-overview · openai.com/business/pricing Checked
Models per plan
Which Codex models each plan lists. Plans with no model list are shown as not published.
| Plan | Models in Codex | Default |
|---|---|---|
| Free | GPT-6 Luna at Standard speed in the desktop app, subject to rollout | GPT-6 Luna (only model listed for Free) |
| Go | GPT-6 Luna at Standard speed in the desktop app, subject to rollout | GPT-6 Luna (only model listed for Go) |
| Plus | GPT-6.1 Sol and GPT-6 Luna (plus GPT-6 Sol, GPT-6 Astra limited, GPT-5.6 family during rollout) | GPT-6.1 Sol when available (recommended) |
| Pro 100 | Not published by OpenAI | Not published by OpenAI |
| Pro 200 | Not published by OpenAI | Not published by OpenAI |
| Pro 500 | Not published by OpenAI | Not published by OpenAI |
| Business Standard seat | Not published by OpenAI | Not published by OpenAI |
| Business Premium seat | Not published by OpenAI | Not published by OpenAI |
| Business Codex seat | Not published by OpenAI | Not published by OpenAI |
| Enterprise / Edu | Not published by OpenAI | Not published by OpenAI |
| API key | follows API models available to your key | Not published by OpenAI |
Sources developers.openai.com/codex/pricing · developers.openai.com/codex/models Checked
5-hour message ranges
Local messages per 5-hour window on Plus (Business Standard uses the same estimates). OpenAI publishes these as ranges because usage depends on task size. Log scale.
GPT-6 Astra
5–45
GPT-6.1 Sol
15–160
GPT-6 Sol
15–150
GPT-6 Luna
350–3,000
- Weekly limits may also apply. Both windows need allowance: a new 5h window starts with your first message after the previous window ends.
- Pro plans: Pro plans currently have no five-hour limit; weekly limits may apply. Business Premium: 5x Standard usage, no 5-hour limit.
- Local messages and cloud chats share the allowance; cloud tasks may use more.
- Free and Go: Limited. Numbers Not published by OpenAI.
Sources developers.openai.com/codex/pricing · help.openai.com/en/articles/20001516-managing-usage-with-gpt-6-astra-in-work-and-codex Checked
Fast & Ultrafast
Faster modes use your included allowance, and credits, faster than Standard (1×).
Standard
Included usage1×
Credit rate1×
Fast
Included usage2.5×
Credit rate2×
Ultrafast (Astra, 6.1 Sol)
Included usage8×
Credit rate6×
- Ultrafast plans: Included (GPT-6 Astra and GPT-6.1 Sol; 8x included usage, 6x credit rate) on Pro 500; Pro 100 and Pro 200 not included.
- Image generation: Uses included limits ~3-5x faster; not on Free.
Sources developers.openai.com/codex/pricing · help.openai.com/en/articles/9793128-what-is-chatgpt-pro Checked
What happens at the limit
- 1
Active turn may finish (fair use)
- 2
Plus/Pro can buy credits
- 3
Business/Edu/Enterprise flexible pricing buy workspace credits
- 4
Use an API key at API rates
- 5
Switch to a lower-cost model
- In ChatGPT
- At a model limit it may be unavailable until its allowance resets; reset time shown when available; use another available model
- Automatic fallback model
- Not published by OpenAIno automatic fallback model stated; pages advise switching to a lower-cost model
- Price of a credit
- Not published by OpenAIpurchase amounts and prices vary by account, region and plan
- Can support reset it?
- Support cannot reset ChatGPT or Codex limits
- Where to check usage
- Usage dashboard; /status in Codex CLI
- Banked resets
- Rate-limit reset banking for Plus and Pro (one free reset at launch; referrals earn more) added 2026-06-11; banked resets granted Sep 3 and Sep 4 2026; automatic reset Sep 7 2026
Credit rates per 1M tokens7
| Model | Input | Cached input | Output |
|---|---|---|---|
| GPT-6 Astra | 250 cr | 25 cr | 1250 cr |
| GPT-6.1 Sol | 50 cr | 2.5 cr | 250 cr |
| GPT-6 Sol | 50 cr | 5 cr | 250 cr |
| GPT-6 Luna | 2.5 cr | 0.25 cr | 12.5 cr |
| GPT-5.6 Sol | 100 cr | 10 cr | 500 cr |
| GPT-5.6 Terra | 50 cr | 5 cr | 300 cr |
| GPT-5.6 Luna | 5 cr | 0.5 cr | 30 cr |
Sources help.openai.com/en/articles/20001516-managing-usage-with-gpt-6-astra-in-work-and-codex · developers.openai.com/codex/pricing · help.openai.com/en/articles/9793128-what-is-chatgpt-pro · help.openai.com/en/articles/12642688-using-credits-for-flexible-usage-in-chatgpt-personal-plans Checked
Surfaces
Where you can run Codex, and which accounts each one accepts.
- ChatGPT desktop appCodex merged in 2026-07-09, macOS/Windows; Linux preview 2026-08-11
Plus/Pro/Business/Enterprise/API key
- Codex CLI
all paid + API key
- IDE extension
all paid + API key
- Codex cloud / web
paid ChatGPT plans, not API key
- GitHub code review and @codex PR delegation
paid ChatGPT plans, not API key
- Codex SDK / codex exec
all paid + API key
- Slack / Linear integrations
paid ChatGPT plans
- iOS / mobile remote
paid ChatGPT plans; Codex Remote GA 2026-06-25
Source developers.openai.com/codex/pricing Checked
Features by plan
yes/no read from the pricing page matrix; empty cells on the page interpreted as yes, — as no
| Feature | Plus | Pro | Business | Enterprise/Edu | API key |
|---|---|---|---|---|---|
| Codex cloud | ✓ | ✓ | ✓ | ✓ | ✕ |
| Codex CLI | ✓ | ✓ | ✓ | ✓ | ✓ |
| IDE extension | ✓ | ✓ | ✓ | ✓ | ✓ |
| Codex access tokens | ✕ | ✕ | ✓ | ✓ | ✕ |
| Ultrafast | ✕ | Pro $500 only | ✕ | eligible | API |
| Composer predictions (beta) | ✕ | limited | ✕ | ✕ | ✕ |
| Computer Use | limited (region) | limited (region) | limited (region) | limited (region) | limited (region) |
| Connectors | ✓ | ✓ | ✓ | ✓ | ✕ |
| Memories | limited (region) | limited (region) | limited (region) | limited (region) | limited (region) |
| Computer History | ✕ | ✓ | ✓ | ✓ | ✕ |
| Sites | ✓ | ✓ | ✓ | ✓ | ✕ |
| GitHub code review | ✓ | ✓ | ✓ | ✓ | ✕ |
| SAML SSO | ✕ | ✕ | ✓ | ✓ | ✕ |
| Compliance API / audit logs | ✕ | ✕ | ✕ | ✓ | ✕ |
✓ included ◐ limited / conditional ✕ not included Not published not stated by OpenAI
Source developers.openai.com/codex/pricing Checked
API pricing
USD per 1M tokens, Standard, short context ≤272K. Applies when Codex runs on an API key.
| Model | Input | Cached input | Cache write | Output |
|---|---|---|---|---|
| gpt-6-astra | $10 | $1 | $12.5 | $50 |
| gpt-6.1-sol | $2 | $0.1 | $2.5 | $10 |
| gpt-6-sol | $2 | $0.2 | $2.5 | $10 |
| gpt-6-luna | $0.1 | $0.01 | $0.125 | $0.5 |
| gpt-5.6-sol | $4 | $0.4 | $5 | $20 |
| gpt-5.6-terra | $2 | $0.2 | $2.5 | $12 |
| gpt-5.6-luna | $0.2 | $0.02 | $0.25 | $1.2 |
| gpt-5.3-codex | $1.75 | $0.175 | Not captured | $14 |
- Batch and Flex = 50% of Standard
- Fast = 2x
- Ultrafast gpt-6-astra $60/$6/$75/$300, gpt-6.1-sol $12/$0.60/$15/$60
- Data-residency endpoints +10% uplift
- GPT-5.6 Sol promo pricing through at least 2026-11-21
Long-context prices (over 272K)
| Model | Input | Cached | Cache write | Output |
|---|---|---|---|---|
| gpt-6-astra | $20 | $2 | $25 | $75 |
| gpt-6.1-sol | $4 | $0.2 | $5 | $15 |
| gpt-6-sol | Not captured | Not captured | Not captured | Not captured |
| gpt-6-luna | Not captured | Not captured | Not captured | Not captured |
| gpt-5.6-sol | Not captured | Not captured | Not captured | Not captured |
| gpt-5.6-terra | Not captured | Not captured | Not captured | Not captured |
| gpt-5.6-luna | Not captured | Not captured | Not captured | Not captured |
| gpt-5.3-codex | Not captured | Not captured | Not captured | Not captured |
"Not captured" means the price was not on the OpenAI page as we checked it.
Source developers.openai.com/api/docs/pricing Checked
What changed
Dated entries from OpenAI's Codex changelog.
GPT-5.5 retired from Codex with ChatGPT sign-in
Composer predictions beta (Pro)
GPT-6.1 Sol Ultrafast (Pro $500, eligible Enterprise/Edu)
GPT-6.1 Sol in Codex and ChatGPT Work
GPT-6 Sol and GPT-6 Luna rolling out; Free/Go get Luna in desktop app
GPT-5.3-Codex-Spark deprecated
Earlier changes9
codex mcp-server removed
gpt-5.4 and gpt-5.4-mini retired from Codex (ChatGPT sign-in)
GitLab support in Codex cloud (beta)
ChatGPT desktop app Linux preview; import from Claude Code/Cursor
Codex joins the ChatGPT desktop app
Codex Remote GA
Rate-limit reset banking for Plus and Pro
Sites preview in Codex app
Codex with Amazon Bedrock
Source developers.openai.com/codex/changelog Checked
FAQ
Why do I hit Codex limits so fast?
Usage depends on model, context, reasoning, tool use, retrieval and caching; Astra uses allowance faster; Fast mode 2.5x and image gen 3-5x faster. Use a smaller model, trim AGENTS.md and MCP servers.
developers.openai.com/codex/pricing · checked
Does Codex finish my task when I hit the limit?
If you reach limits during an active turn, the agent can continue that turn, subject to fair use limits.
developers.openai.com/codex/pricing · checked
Is GPT-6 Astra included on Plus?
Plus includes limited Astra usage in Work and Codex, with optional credits; Pro $100/$200 and Business Premium can use their full allowance for Astra. GPT-6 Pro in Chat is not on Plus.
help.openai.com/en/articles/20001275-chatgpt-work-and-codex · checked
What is Sol 6.1?
GPT-6.1 Sol (gpt-6.1-sol), released in Codex 2026-09-29: near-Astra performance at lower cost; Plus and above; not in Chat.
developers.openai.com/codex/models · checked
Why did my Pro 200 allowance drop?
New non-grandfathered Pro 200 subscriptions get a lower allowance; grandfathered users (active Pro 200 between 2026-09-22 and 2026-09-29 10am PT) keep the previous allowance through 2026-10-29. Price stays $200.
help.openai.com/en/articles/9793128-what-is-chatgpt-pro · checked
Can I buy more usage?
Yes, credits for eligible features (Codex, ChatGPT Work, Word/Excel/PowerPoint); included usage first, then credits. Prices vary by account/region/plan.
help.openai.com/en/articles/9793128-what-is-chatgpt-pro · checked
Can Support reset my limits?
No. Support cannot reset ChatGPT or Codex limits.
help.openai.com/en/articles/9793128-what-is-chatgpt-pro · checked
Is there annual billing for Pro?
No, Pro is billed monthly only.
help.openai.com/en/articles/9793128-what-is-chatgpt-pro · checked
Can I use Codex free?
Free and Go get GPT-6 Luna at Standard speed in the desktop app, subject to rollout; limited Codex access.
developers.openai.com/codex/pricing · checked
Does API-key Codex use my subscription?
No. With an API key you pay API token prices; no cloud features.
developers.openai.com/codex/pricing · checked
Sources developers.openai.com/codex/pricing · help.openai.com/en/articles/20001275-chatgpt-work-and-codex · developers.openai.com/codex/models · help.openai.com/en/articles/9793128-what-is-chatgpt-pro Checked
Regions: Not published by OpenAIcountry availability list not checked); some features region-limited (Computer Use, Memories, Record & Replay, Computer History developers.openai.com/codex/pricing