Plans

Also included

• Codex / Work share one 5h + weekly pool
• OpenAI publishes 5h local ranges, not a hard message cap

Also included

• Deep research included
• Some models still carry their own caps

Also included

• New $200 sign-ups paused since 10 Sep 2026
• Existing $200 renewals continue

Also included

• Fable 5.1 is billed from usage credits. Fable 5 uses the same rule.
• Code work shares the same rolling 5h + weekly cap
• Annual bill is $17 / mo ($200 up front)

Also included

• Weekly cap still shared across models
• Fable 5.1 uses 50% of the weekly limits. Fable 5 uses the same rule.

Also included

• Weekly cap still shared across models
• Fable 5.1 uses 50% of the weekly limits. Fable 5 uses the same rule.

How the two pools work

• Composer 2.5 and Grok 4.5, 4.6, and 4.7, including Fast, share the own pool.
• Own pool can spend leftover Other. Other cannot use the own pool.
• Both reset on the billing date, not the 1st of the month.
• Live docs print Included for both. Own-pool dollars are dashboard observations.

Also included

• Two monthly pools: Composer/Grok and Other Models
• Own pool can spill into leftover Other; Other cannot use the own pool
• Resets on the subscription billing date, not the 1st

How the two pools work

• Composer 2.5 and Grok 4.5, 4.6, and 4.7, including Fast, share the own pool.
• Own pool can spend leftover Other. Other cannot use the own pool.
• Both reset on the billing date, not the 1st of the month.
• Live docs print Included for both. Own-pool dollars are dashboard observations.

Also included

• Same model set as Pro, including Grok 4.7 in the Cursor Models pool
• Own pool can spill into leftover Other; Other cannot use the own pool
• Resets on the subscription billing date, not the 1st

How the two pools work

• Composer 2.5 and Grok 4.5, 4.6, and 4.7, including Fast, share the own pool.
• Own pool can spend leftover Other. Other cannot use the own pool.
• Both reset on the billing date, not the 1st of the month.
• Live docs print Included for both. Own-pool dollars are dashboard observations.
• Some Ultra dashboards later showed $500 Other; the last official Other figure stays $400.

Also included

• Same model set as Pro, including Grok 4.7 in the Cursor Models pool
• Own pool can spill into leftover Other; Other cannot use the own pool
• Resets on the subscription billing date, not the 1st

How the Gemini pools work

• Gemini app, Antigravity, AI Studio, Jules, Flow, and Cloud credits are separate pools. They do not spill into each other.
• The app and Antigravity refresh every 5 hours until a weekly cap. App usage is compute-based; there is no published dollar or turn count.
• Antigravity CLI and the desktop app share that same baseline pool. Antigravity numbers are official relative multipliers — 1× / 5× Pro / 20× Pro — not a published dollar window.
• AI Studio included quota is daily and UI-only. API keys and external apps bill separately.
• Cloud credits are official Developer Program dollars for Cloud / Gemini API after billing is on. They are not Google One AI credits and do not fund the Gemini app.
• The old Gemini CLI daily request counts (Pro 1,500 / Ultra 2,000) stopped serving individual Google AI accounts on 18 Jun 2026. Enterprise Code Assist and API-key Gemini CLI remain a different daily pool.
• Jules is official: Pro 100 tasks and 15 concurrent per rolling 24h; Ultra 300 and 60. Ultra 5x and 20x share the Ultra Jules row.
• Flow video is 1,000 / 10,000 / 25,000 credits a month; Flow Music is 10,000 / 30,000 / 30,000. Those are not API dollars.

Also included

• 3.8 Flash in Gemini, Search AI Mode, and Sheets (official blog, 2 Sep 2026)
• Plan page still lists Gemini 3.1 Pro, Deep Research, and Omni Flash
• Gemini Spark and Chrome auto browse (US)
• YouTube Premium Lite, Health Premium, Home Premium Standard
• $10 / mo Google Cloud credits; Google Flow video 1,000 / mo; Flow Music 10,000 / mo
• Jules 100 tasks / day, 15 concurrent (official, rolling 24h)
• Expanded AI Studio and Antigravity limits; those pools do not share with the app

How the Gemini pools work

• Gemini app, Antigravity, AI Studio, Jules, Flow, and Cloud credits are separate pools. They do not spill into each other.
• The app and Antigravity refresh every 5 hours until a weekly cap. App usage is compute-based; there is no published dollar or turn count.
• Antigravity CLI and the desktop app share that same baseline pool. Antigravity numbers are official relative multipliers — 1× / 5× Pro / 20× Pro — not a published dollar window.
• AI Studio included quota is daily and UI-only. API keys and external apps bill separately.
• Cloud credits are official Developer Program dollars for Cloud / Gemini API after billing is on. They are not Google One AI credits and do not fund the Gemini app.
• The old Gemini CLI daily request counts (Pro 1,500 / Ultra 2,000) stopped serving individual Google AI accounts on 18 Jun 2026. Enterprise Code Assist and API-key Gemini CLI remain a different daily pool.
• Jules is official: Pro 100 tasks and 15 concurrent per rolling 24h; Ultra 300 and 60. Ultra 5x and 20x share the Ultra Jules row.
• Flow video is 1,000 / 10,000 / 25,000 credits a month; Flow Music is 10,000 / 30,000 / 30,000. Those are not API dollars.

Also included

• 3.8 Flash in Gemini, Search AI Mode, and Sheets (official blog, 2 Sep 2026)
• Deep Think, Gemini Agent (US, English), Project Genie
• YouTube Premium individual and Home Premium Advanced
• $40 / mo Google Cloud credits; Google Flow video 10,000 / mo; Flow Music 30,000 / mo
• Jules 300 tasks / day, 60 concurrent (official, rolling 24h)
• Higher AI Studio and Antigravity limits; those pools do not share with the app

How the Gemini pools work

• Gemini app, Antigravity, AI Studio, Jules, Flow, and Cloud credits are separate pools. They do not spill into each other.
• The app and Antigravity refresh every 5 hours until a weekly cap. App usage is compute-based; there is no published dollar or turn count.
• Antigravity CLI and the desktop app share that same baseline pool. Antigravity numbers are official relative multipliers — 1× / 5× Pro / 20× Pro — not a published dollar window.
• AI Studio included quota is daily and UI-only. API keys and external apps bill separately.
• Cloud credits are official Developer Program dollars for Cloud / Gemini API after billing is on. They are not Google One AI credits and do not fund the Gemini app.
• The old Gemini CLI daily request counts (Pro 1,500 / Ultra 2,000) stopped serving individual Google AI accounts on 18 Jun 2026. Enterprise Code Assist and API-key Gemini CLI remain a different daily pool.
• Jules is official: Pro 100 tasks and 15 concurrent per rolling 24h; Ultra 300 and 60. Ultra 5x and 20x share the Ultra Jules row.
• Flow video is 1,000 / 10,000 / 25,000 credits a month; Flow Music is 10,000 / 30,000 / 30,000. Those are not API dollars.

Also included

• 3.8 Flash in Gemini, Search AI Mode, and Sheets (official blog, 2 Sep 2026)
• Highest access to Deep Think, Agent, Project Genie, and Google Flow
• YouTube Premium individual and Home Premium Advanced
• $100 / mo Google Cloud credits; Google Flow video 25,000 / mo; Flow Music 30,000 / mo
• Jules 300 tasks / day, 60 concurrent (official, rolling 24h)
• Highest AI Studio and Antigravity limits; those pools do not share with the app

How the SuperGrok pool works

• Chat, Imagine, Voice, and Build share one weekly compute pool. The Usage tab also breaks out API on that same bar.
• The meter is a percentage. There is no published dollar window or turn count. Chat costs little; video and long coding cost more.
• The pool resets weekly on the date in Settings → Usage, not a fixed weekday for everyone.
• Extra Usage Credits start at $5 on the web, spend after the included pool, expire in one year, and bill at standard rates.
• After the weekly cap, free-tier Chat and Voice remain on their own schedule.
• Grok Bot is not grok.com. It has its own weekly grant. SuperGrok / Plus / Heavy can link that grant; Bot does not spend the grok.com pool. If you also have Cursor, xAI says Bot uses whichever grant is larger.
• console.x.ai prepaid API credits are a separate team billing system, not SuperGrok Extra Usage Credits.
• 720p video falls back to 480p after the tier cap. Public Heavy cards still omit the monthly dollar; third-party $300 figures are not used here.
• Grok 4.7 launched 21 Sep 2026 on Cursor, Grok Build, and the API. grok.com, the mobile apps, and Grok in X still publish Grok 4.6.

Also included

• Grok Bot access as a separate weekly grant, not the grok.com pool
• Image and video generation
• Chat, Imagine, Voice, and Build share one weekly percent pool
• Extra Usage Credits from $5 after the pool (web, expire in one year)
• SOC 2 Type I & II

How the SuperGrok pool works

• Chat, Imagine, Voice, and Build share one weekly compute pool. The Usage tab also breaks out API on that same bar.
• The meter is a percentage. There is no published dollar window or turn count. Chat costs little; video and long coding cost more.
• The pool resets weekly on the date in Settings → Usage, not a fixed weekday for everyone.
• Extra Usage Credits start at $5 on the web, spend after the included pool, expire in one year, and bill at standard rates.
• After the weekly cap, free-tier Chat and Voice remain on their own schedule.
• Grok Bot is not grok.com. It has its own weekly grant. SuperGrok / Plus / Heavy can link that grant; Bot does not spend the grok.com pool. If you also have Cursor, xAI says Bot uses whichever grant is larger.
• console.x.ai prepaid API credits are a separate team billing system, not SuperGrok Extra Usage Credits.
• 720p video falls back to 480p after the tier cap. Public Heavy cards still omit the monthly dollar; third-party $300 figures are not used here.
• Grok 4.7 launched 21 Sep 2026 on Cursor, Grok Build, and the API. grok.com, the mobile apps, and Grok in X still publish Grok 4.6.

Also included

• Significantly higher usage across Chat, Imagine, Voice, and Build
• Lightning-fast replies and priority at peak times
• Early access to new features
• Larger linked Grok Bot weekly grant than SuperGrok

How the SuperGrok pool works

• Chat, Imagine, Voice, and Build share one weekly compute pool. The Usage tab also breaks out API on that same bar.
• The meter is a percentage. There is no published dollar window or turn count. Chat costs little; video and long coding cost more.
• The pool resets weekly on the date in Settings → Usage, not a fixed weekday for everyone.
• Extra Usage Credits start at $5 on the web, spend after the included pool, expire in one year, and bill at standard rates.
• After the weekly cap, free-tier Chat and Voice remain on their own schedule.
• Grok Bot is not grok.com. It has its own weekly grant. SuperGrok / Plus / Heavy can link that grant; Bot does not spend the grok.com pool. If you also have Cursor, xAI says Bot uses whichever grant is larger.
• console.x.ai prepaid API credits are a separate team billing system, not SuperGrok Extra Usage Credits.
• 720p video falls back to 480p after the tier cap. Public Heavy cards still omit the monthly dollar; third-party $300 figures are not used here.
• Grok 4.7 launched 21 Sep 2026 on Cursor, Grok Build, and the API. grok.com, the mobile apps, and Grok in X still publish Grok 4.6.

Also included

• Highest linked Grok Bot weekly grant
• Checkout is the Get Heavy button on x.ai/pricing
• Public cards still omit the monthly dollar — third-party $300 figures are not used here