Codex Plus has one shared five-hour allowance, with official approximate local-message ranges by model. Pro plans ($100, $200 or $500 per month) currently have no five-hour limit. Weekly limits may apply to both. Credits can continue after included usage.
Last verified · Codex
Usage evidence reviewed . Other product facts retain their earlier verification date.
Fact Plus $20.00/month; Pro from $100.00/month (Plans at $100, $200, or $500 per month).
Plus five-hour quota
Fact Shared five-hour local-message estimates by model, for example 15–160 with GPT-6.1 Sol. These are not fixed limits.
Pro five-hour quota
Fact Pro plans currently have no five-hour limit.
Weekly window
Fact Additional weekly limits may apply. Unknown OpenAI does not publish one fixed weekly message count.
Reset
Fact Remaining allowance and reset time appear in the Codex usage dashboard and CLI /status. TraceCheck does not invent a UTC clock.
Shared pool
Fact Local messages and cloud chats share the plan allowance. Switching models does not create a new pool.
After the limit
Fact Plus and Pro can buy credits after included usage. Compare waiting, a banked reset, credits, and upgrading in the reset vs credits guide.
Last verified
Oct 11, 2026, 9:30 AM UTC
Changed since September
What changed in Codex limits and models
Fact Codex now recommends GPT-6.1 Sol and adds GPT-6 Sol and GPT-6 Luna. GPT-5.3-Codex-Spark retired on September 14, and GPT-5.5 retires from ChatGPT-signed-in Codex on October 14, 2026. Verified Oct 11, 2026, 9:30 AM UTC.
Fact Codex Plus limits now list the GPT-6 models: GPT-6.1 Sol 15–160, GPT-6 Sol 15–150 and GPT-6 Luna 350–3,000 local messages per five hours. GPT-5.6 Terra has no row. Verified Oct 11, 2026, 9:30 AM UTC.
Fact Codex Pro is now three plans ($100, $200 and $500 per month), and Pro plans currently have no five-hour limit. The Pro 5x and 20x message columns are gone from the usage table. Verified Oct 11, 2026, 9:30 AM UTC.
If GPT-5.5 is saved in config.toml, a custom agent or a script, switch it before October 14, 2026: GPT-6 Sol on Plus, Pro, Business, Enterprise and Edu; GPT-6 Luna on Free and Go. API-key use is not affected. Step-by-step: Codex models and the GPT-5.5 replacement.
Plus vs Pro at a glance
Plan facts from the official pricing and usage pages. Model ranges are not collapsed into one number.
Plan
Subscription price
Capacity rule
Single numeric allowance
Plus
Fact $20.00/month
Shared five-hour usage; weekly limits may also apply.
Unknown as one number Official approximate ranges vary by model in the Plus column.
Pro
Fact From $100.00/month
Pro plans currently have no five-hour limit; weekly limits may apply. Choose a $100, $200 or $500 plan; Ultrafast mode is on Pro $500.
Fact No five-hour limit currently. Unknown Weekly allowance not published.
Each plan below is a pool graph: consumers drain a pool, a reset clock refills it, and exhaustion says what happens at empty. Numeric capacity is shown only when an official source publishes it for that exact pool.
No single Plus five-hour quantity is published; official ranges vary by model.
Consumer · surface Codex local — Fact Local messages and cloud chats share the Plus five-hour allowance.
Consumer · surface Codex cloud — Fact Cloud chats share the Plus allowance and may use more than local messages.
Consumer · surface ChatGPT Work — Fact ChatGPT Work shares Codex usage where the plan supports it.
Consumer · model GPT-6.1 Sol — Fact Official Plus-column approximate local messages / 5h: 15–160. Not a fixed limit.
Consumer · model GPT-6 Astra — Fact Official Plus-column approximate local messages / 5h: 5–45. Not a fixed limit.
Consumer · model GPT-6 Sol — Fact Official Plus-column approximate local messages / 5h: 15–150. Not a fixed limit.
Consumer · model GPT-6 Luna — Fact Official Plus-column approximate local messages / 5h: 350–3,000. Not a fixed limit.
Consumer · model GPT-5.6 Sol — Fact Official Plus-column approximate local messages / 5h: 10–100. Not a fixed limit.
Consumer · model GPT-5.6 Terra
Consumer · model GPT-5.6 Luna — Fact Official Plus-column approximate local messages / 5h: 250–2,000. Not a fixed limit.
Reset · rollingFact Remaining allowance and reset time are shown in the Codex usage dashboard and CLI /status. TraceCheck does not invent a UTC clock. Visible in Codex usage dashboard, CLI /status.
At limit · credits, wait for reset, upgradeFact Plus users can buy credits, wait for the displayed reset, switch to a lower-cost model, or upgrade. The current turn may finish subject to fair use.
Pro plans currently have no five-hour limit. Weekly limits may apply; the numeric weekly cap is not publicly disclosed.
Consumer · surface Codex local — Fact Local messages and cloud chats share the Pro usage allowance. Pro plans currently have no five-hour limit.
Consumer · surface Codex cloud
Consumer · model GPT-6.1 Sol
Consumer · model GPT-6 Astra
Consumer · model GPT-6 Sol
Consumer · model GPT-6 Luna
Consumer · model GPT-5.6 Sol
Consumer · model GPT-5.6 Terra
Consumer · model GPT-5.6 Luna
Reset · weeklyUnknown Weekly limits may apply. No universal weekly reset timestamp is published. Visible in Codex usage dashboard.
At limit · credits, wait for reset, upgradeFact Pro users who reach their usage limit can buy credits, wait for the displayed reset, switch model, or move to a larger Pro plan.
DocumentedShared pools: Plus shared five-hour allowance; Plus weekly limit; Pro weekly limit. Only documented consumers on those pools draw the same allowance. A surface not listed as a consumer is not assumed to share it.
PartialRemaining allowance and reset time are shown in the Codex usage dashboard and CLI /status. TraceCheck does not invent a UTC clock. Weekly limits may apply. No universal weekly reset timestamp is published. Weekly limits may apply. No universal weekly reset timestamp is published.
Mechanics
rolling: Remaining allowance and reset time are shown in the Codex usage dashboard and CLI /status. TraceCheck does not invent a UTC clock. Visible in Codex usage dashboard, CLI /status.
weekly: Weekly limits may apply. No universal weekly reset timestamp is published. Visible in Codex usage dashboard.
weekly: Weekly limits may apply. No universal weekly reset timestamp is published. Visible in Codex usage dashboard.
Next action: Look in Codex usage dashboard, CLI /status.
Unknown / conflict
Unknown No single Plus five-hour quantity is published; official ranges vary by model.
Unknown OpenAI says weekly limits may apply but does not publish one fixed weekly message count.
Unknown Pro plans currently have no five-hour limit. Weekly limits may apply; the numeric weekly cap is not publicly disclosed.
Fact A rolling five-hour allowance is shared across local messages and cloud chats where applicable.
Fact Additional weekly limits may apply even when a short-window allowance remains.
Fact Check the Codex usage dashboard or CLI /status for your remaining allowance and reset time.
Unknown These sources do not establish one universal reset day or hour for every account.
Derived Do not multiply one model's message estimate by the number of models or by five-hour periods to invent a daily or monthly quota.
What happens after included usage?
Fact Additional credits can extend usage after included limits; credit consumption varies by model and speed.
Fact If you reach a limit during an active turn, the agent may finish that turn subject to fair use. Plus and Pro can purchase credits without changing plans. Credits extend usage; they are not a usage-window reset. For the wait / banked-reset / credits / upgrade choice, use the reset vs credits guide.
Derived Higher included capacity and additional credits address different needs: moving to Pro removes the five-hour limit, while credits extend usage beyond whatever limit you hit. These facts cannot predict completed coding tasks per dollar.
Model-specific five-hour estimates
Fact OpenAI publishes approximate local messages per five-hour period for Plus and Standard Business. They are not fixed message limits. Cloud tasks may use more of the allowance than local messages.
Official columns only. Pro has no column because Pro plans currently have no five-hour limit.
Model
Plus
Standard Business
Pro
GPT-6.1 Sol
Fact 15–160 local messages / 5h
Fact 15–160 local messages / 5h
No five-hour limit
GPT-6 Astra
Fact 5–45 local messages / 5h
Fact 5–45 local messages / 5h
No five-hour limit
GPT-6 Sol
Fact 15–150 local messages / 5h
Fact 15–150 local messages / 5h
No five-hour limit
GPT-6 Luna
Fact 350–3,000 local messages / 5h
Fact 350–3,000 local messages / 5h
No five-hour limit
GPT-5.6 Sol
Fact 10–100 local messages / 5h
Fact 10–100 local messages / 5h
No five-hour limit
GPT-5.6 Luna
Fact 250–2,000 local messages / 5h
Fact 250–2,000 local messages / 5h
No five-hour limit
Fact GPT-5.6 Terra remains selectable during the GPT-6 rollout but has no row in the current table. GPT-5.3-Codex-Spark retired on September 14, 2026.
Not directly comparable These ranges are approximate local messages, not a universal token, credit, or task score. Model rows share one five-hour allowance; they are not separate pools you can add together.
What is not disclosed
Unknown OpenAI does not publish a fixed weekly numeric cap, a universal reset clock, a Pro weekly allowance, or a single token conversion of the subscription pool. Free and Go have no model-specific message table.
Free and Go have no model-specific message table in the verified usage chart. A subscription price is not an API token price. Unknown — not publicly disclosed as a single weekly number.
The chatgpt.com table still lists GPT-5.5, GPT-5.4, and GPT-5.4 mini rows. GPT-5.4 and GPT-5.4 mini retired from ChatGPT-signed-in Codex on August 31, 2026, and GPT-5.5 retires on October 14, 2026, so those rows are not treated as current availability.
Codex CLI pricing versus API token cost
Fact Plus includes the CLI alongside other Codex interfaces. Separately configured API/provider access has its own usage billing and limits; the ChatGPT subscription amount is not an API token price.
Unknown No single token cost covers every model and access path. This page therefore does not convert a subscription allowance to tokens or quote an unverified API rate.
Official sources conflict
Official sources conflict Official sources conflict. The Codex pricing table still displays legacy GPT-5.4 and GPT-5.4 mini rows, while the Help Center states those models are no longer available in Codex for users signed in with ChatGPT from August 31, 2026.
Codex now recommends GPT-6.1 Sol and adds GPT-6 Sol and GPT-6 Luna. GPT-5.3-Codex-Spark retired on September 14, and GPT-5.5 retires from ChatGPT-signed-in Codex on October 14, 2026. — Codex models on Plus: GPT-6 Astra, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna; Pro also GPT-5.3-Codex-Spark → GPT-6.1 Sol (recommended), GPT-6 Astra, GPT-6 Sol and GPT-6 Luna, with GPT-5.6 Sol, Terra and Luna available during the rollout. Spark retired September 14, 2026; GPT-5.5 retires October 14, 2026. Verified Oct 11, 2026, 9:30 AM UTC.
Codex Plus limits now list the GPT-6 models: GPT-6.1 Sol 15–160, GPT-6 Sol 15–150 and GPT-6 Luna 350–3,000 local messages per five hours. GPT-5.6 Terra has no row. — Plus approximate local messages per five hours: GPT-6 Astra 5–45, GPT-5.6 Sol 10–100, GPT-5.6 Terra 25–200, GPT-5.6 Luna 250–2,000 → Plus and Standard Business approximate local messages per five hours: GPT-6.1 Sol 15–160, GPT-6 Astra 5–45, GPT-6 Sol 15–150, GPT-6 Luna 350–3,000, GPT-5.6 Sol 10–100, GPT-5.6 Luna 250–2,000. Verified Oct 11, 2026, 9:30 AM UTC.
Codex Pro is now three plans ($100, $200 and $500 per month), and Pro plans currently have no five-hour limit. The Pro 5x and 20x message columns are gone from the usage table. — Pro was one plan from $100/month with official 5x and 20x capacity columns of approximate local messages per five hours (GPT-6 Astra 25–225 at 5x, 100–900 at 20x) → Pro plans at $100, $200 or $500 per month. Pro plans currently have no five-hour limit; weekly limits may apply. Ultrafast mode is on Pro $500. Verified Oct 11, 2026, 9:30 AM UTC.
Codex Plus approximate five-hour local-message ranges re-verified from the official Plus column; they are not a single collapsed quota. — Unknown — current model-specific message estimates require re-verification. A shared five-hour allowance applies, and additional weekly limits may apply. → Approximate official local-message estimates per five-hour period from the Plus column: GPT-6 Astra 5–45; GPT-5.6 Sol 10–100; GPT-5.6 Terra 25–200; GPT-5.6 Luna 250–2,000. Shared pool; weekly limits may apply; estimates are not fixed limits.. Verified Sep 7, 2026, 4:00 PM UTC.