Save money › AI usage limits
AI usage limits: which plan do you actually need?
Paying for API tokens instead of a chat plan? See OpenAI vs Claude vs Gemini $/MTok for your volume: AI API cost calculator →
Message caps, 5-hour windows and weekly limits, explained for 8 AI tools. Pick a tool and how much you use it to see the plan to buy, when the limit resets and what to do when you hit it. From each tool's own pricing and help pages, checked 2026-11-03.
Claude · every day for work
Pro · $20/mo ($17/mo billed yearly)
Claude's own page calls Pro the plan “for everyday work”.
- What you get: At least 5x more usage per 5-hour session than Free, plus a weekly limit; includes Claude Code
- Limit resets: A rolling 5-hour session window on every plan, plus a weekly limit on paid plans. Ultra has no 5-hour limit for verified accounts.
- If you hit the limit: Wait for your limit to reset; move to a higher plan; on paid plans, turn on usage credits to keep working at standard API rates.
- Next step up: Max ($100/mo)
From Claude's own pages, checked 2026-11-03: Pricing and usage FAQ (opens in a new tab) · Claude review
How each tool's limits work, side by side
“Published numbers” means the company states a count; most only state multipliers (“5x Pro”) or labels (“extensive”).
| Tool | How usage is counted | When it resets | Published numbers | Pay for extra? |
|---|---|---|---|---|
| ChatGPT | Separate limits per feature: everyday text chats are unlimited on every plan (abuse guardrails apply); uploads, images, voice, deep research and Thinking have their own limits. Codex and ChatGPT Work share another allowance. | Chat limits show their own reset time. Codex and Work on Plus: a 5-hour window plus a weekly limit. Pro plans: no 5-hour limit in Codex and Work, weekly allowance only. | Plus, Codex/Work estimate per 5 hours: 15–150 GPT-6 Sol messages, 5–45 GPT-6 Astra messages | Credits for Codex, Work and Office add-ins only |
| Claude | One shared pool across Claude on web, desktop, mobile and Claude Code. No fixed message count: it depends on chat length, file size and the model you pick. | A rolling 5-hour session window on every plan, plus a weekly limit on paid plans. Ultra has no 5-hour limit for verified accounts. | Multipliers only: Pro at least 5x Free per 5-hour session; Max 5x Pro; Ultra 10x Pro per week | Yes, usage credits at API rates (paid plans) |
| Google Gemini | Compute-based: a limit that depends on how complex your prompt is, the model and features you use, and how long the chat is. Deep Research and video use the most. | Refreshes every 5 hours until you reach your weekly limit. | Multipliers only: AI Plus 2x, AI Pro 4x the no-plan limit; AI Ultra 5x or 20x AI Pro | No paid top-up mentioned for the Gemini app |
| Microsoft Copilot | Per-feature limits within a time period, plus monthly AI credits for some features. Chat is “extensive usage” on every paid plan. | AI credits are a monthly allotment. Feature limits reset after a time period Microsoft doesn't state. | Labels only: Standard / More / Extensive / Highest usage | Not stated |
| Le Chat (Mistral) | Caps on messages, web searches, image generations and coding sessions, higher on each plan. | Some limits are daily (for example Flash answers); others aren't stated. | Flash answers: 150 a day on Pro, 200 a day on Team; Pro up to 40x more image generations than Free | Not stated |
| Cursor | One pool of AI usage shared across Cursor and Grok; each product uses it at a different rate. Third-party models are billed separately at API list price. | Your allowance refreshes weekly. | Multipliers only: Premium 4x Plus, Super 15x, Ultra 50x | Yes, top-offs |
| GitHub Copilot | Monthly GitHub AI Credits (1 credit = $0.01); each chat or agent task costs credits by model and tokens. Code completions and next edit suggestions are unlimited on paid plans and use no credits. | Monthly, at 00:00 UTC on the 1st of each calendar month. Unused credits don't carry over. | Free: 2,000 completions a month. Pro 1,500, Pro+ 7,000, Max 20,000 AI credits a month | Yes, $0.01 per AI credit |
| Windsurf | A usage quota per plan; each message costs more or less depending on the model, task size and reasoning. | Refreshes automatically every day and every week. | No numbers: Free “light quota”, Pro “increased quotas”, Max “significantly higher quotas” | Yes, at API pricing |
Every plan's limit, tool by tool
Open a tool to see each plan's price, what it includes and who the company says it's for.
ChatGPT plans and limits
Our pick by use: A few times a week: Free · Every day for work: Plus · All day or with agents: Pro 100 · Open in the picker
| Plan | Price | What the plan includes | ChatGPT says it's for |
|---|---|---|---|
| Free | $0 | Unlimited text chats with GPT-6; limited uploads, image creation, voice, deep research and Codex | “Intelligence for everyday tasks” |
| Go | $8/mo (price checked 2026-10-24) | More messages with tools, more uploads, image creation and voice chats; longer memory; may include ads | “Expanded access” |
| Plus | $20/mo (price checked 2026-10-24) | Advanced reasoning with GPT-6; expanded messages, uploads and deep research; expanded Codex usage (estimate: 15–150 GPT-6 Sol messages per 5 hours) | “Do more with advanced intelligence” |
| Pro 100 | $100/mo | Pro reasoning powered by GPT-6 Astra; longer Codex and ChatGPT Work sessions with no 5-hour limit there; maximum deep research | “Maximize your productivity” |
| Pro 200 | $200/mo | More usage than Pro 100 | — |
| Pro 500 | $500/mo | The highest included usage of the three Pro plans; the only Pro plan with Ultrafast | — |
How it's counted: Separate limits per feature: everyday text chats are unlimited on every plan (abuse guardrails apply); uploads, images, voice, deep research and Thinking have their own limits. Codex and ChatGPT Work share another allowance. Resets: Chat limits show their own reset time. Codex and Work on Plus: a 5-hour window plus a weekly limit. Pro plans: no 5-hour limit in Codex and Work, weekly allowance only. At the limit: ChatGPT may carry on with another available model when you reach a Thinking limit; wait for the reset time ChatGPT shows; buy credits for Codex, ChatGPT Work and the Office add-ins (not every model or feature); upgrade to a plan with more usage. Support can't reset your limits. Fast mode uses your allowance 2.5x faster.
Sources (checked 2026-11-03): Pricing (opens in a new tab) · Models and usage limits (opens in a new tab) · Pro tiers (opens in a new tab) · Work and Codex usage (opens in a new tab) · ToolPick review
Claude plans and limits
Our pick by use: A few times a week: Free · Every day for work: Pro · All day or with agents: Max · Open in the picker
| Plan | Price | What the plan includes | Claude says it's for |
|---|---|---|---|
| Free | $0 | Limited usage every 5 hours and every week | “Try Claude” |
| Pro | $20/mo ($17/mo billed yearly) | At least 5x more usage per 5-hour session than Free, plus a weekly limit; includes Claude Code | “For everyday work” |
| Max | $100/mo | 5x Pro usage every 5 hours and every week; priority access at high traffic times | “For all-day work” |
| Ultra | $200/mo | No 5-hour limit (verified accounts); 10x Pro usage every week | “For your longest tasks” |
How it's counted: One shared pool across Claude on web, desktop, mobile and Claude Code. No fixed message count: it depends on chat length, file size and the model you pick. Resets: A rolling 5-hour session window on every plan, plus a weekly limit on paid plans. Ultra has no 5-hour limit for verified accounts. At the limit: Wait for your limit to reset; move to a higher plan; on paid plans, turn on usage credits to keep working at standard API rates. See where you stand in Settings > Usage.
Sources (checked 2026-11-03): Pricing and usage FAQ (opens in a new tab) · ToolPick review
Google Gemini plans and limits
Our pick by use: A few times a week: No plan (free) · Every day for work: Google AI Pro · All day or with agents: Google AI Ultra 5x · Open in the picker
| Plan | Price | What the plan includes | Google Gemini says it's for |
|---|---|---|---|
| No plan (free) | $0 | Standard limits; 32k-token context window | — |
| Google AI Plus | $4.99/mo | 2x the standard limits; 128k context; Gemini 3.6 Flash and Deep Research | “Essential tools with double the access to Gemini” |
| Google AI Pro | $19.99/mo | 4x the standard limits; 1M-token context; Gemini 4 Argon | — |
| Google AI Ultra 5x | $99.99/mo | 5x AI Pro limits | “5x higher usage limits in Gemini than the AI Pro plan” |
| Google AI Ultra 20x | $199.99/mo | 20x AI Pro limits | — |
How it's counted: Compute-based: a limit that depends on how complex your prompt is, the model and features you use, and how long the chat is. Deep Research and video use the most. Resets: Refreshes every 5 hours until you reach your weekly limit. At the limit: On a Google AI plan, keep chatting with Flash-Lite; wait until your limit refreshes (Gemini tells you when); upgrade to a plan with higher limits. Starting a new chat for a new topic saves usage. Check Settings > Usage Limits.
Sources (checked 2026-11-03): Gemini Apps limits (opens in a new tab) · Google AI plans (opens in a new tab) · ToolPick review
Microsoft Copilot plans and limits
Our pick by use: A few times a week: Copilot (free) · Every day for work: Microsoft 365 Personal · All day or with agents: Microsoft 365 Premium · Open in the picker
| Plan | Price | What the plan includes | Microsoft Copilot says it's for |
|---|---|---|---|
| Copilot (free) | $0 | Limited usage with a free Microsoft account | — |
| Microsoft 365 Personal | $9.99/mo ($99.99/yr) | Advanced Copilot chat and Copilot in Word, Excel, PowerPoint and Outlook with higher usage than free; image generation, voice and vision: standard usage; Researcher and Analyst agents not included | — |
| Microsoft 365 Premium | $199.99/yr ($19.99/mo) (price checked 2026-10-20 (monthly price); $199.99/yr checked 2026-11-03) | Chat: extensive; image generation: extensive; voice and vision: more usage; Researcher and Analyst agents: extensive | “Extensive AI usage and access to exclusive Copilot features” |
How it's counted: Per-feature limits within a time period, plus monthly AI credits for some features. Chat is “extensive usage” on every paid plan. Resets: AI credits are a monthly allotment. Feature limits reset after a time period Microsoft doesn't state. At the limit: Microsoft notifies you in the app; check what's left under Services & subscriptions in your Microsoft account; switch to a plan with higher limits (Microsoft 365 Pro has the highest). AI benefits are for the subscription owner only, even on Family plans.
Sources (checked 2026-11-03): AI credits and limits (opens in a new tab) · Compare Microsoft 365 plans (opens in a new tab) · ToolPick review
Le Chat (Mistral) plans and limits
Our pick by use: A few times a week: Free · Every day for work: Pro · All day or with agents: Pro · Open in the picker
| Plan | Price | What the plan includes | Le Chat (Mistral) says it's for |
|---|---|---|---|
| Free | $0 | Limited messages and web searches; limited coding sessions; basic image generation | — |
| Pro | $14.99/mo (excl. tax) | More messages and web searches; faster responses; 150 Flash answers a day; up to 40x more image generations than Free; all-day coding with Vibe | “Full access to Vibe for all-day coding and long-running tasks” |
| Team | $24.99/user/mo (excl. tax) | Even higher limits; 200 Flash answers a day | “Secure, collaborative workspace for teams building with AI” |
How it's counted: Caps on messages, web searches, image generations and coding sessions, higher on each plan. Resets: Some limits are daily (for example Flash answers); others aren't stated. At the limit: Not stated on the pricing page; upgrade to Pro or Team for higher limits. Verified students get Pro for $5.99/mo.
Sources (checked 2026-11-03): Pricing and FAQ (opens in a new tab) · ToolPick review
Cursor plans and limits
Our pick by use: A few times a week: Free · Every day for work: Premium · All day or with agents: Super · Open in the picker
| Plan | Price | What the plan includes | Cursor says it's for |
|---|---|---|---|
| Free | $0 | Limited agent and chat requests | — |
| Plus | $8/mo | Basic shared usage pool; only Grok models | “Recommended for social creators with limited AI usage” |
| Premium | $30/mo | 4x more usage than Plus; cloud agents, code reviews; third-party models at API list price | “Recommended for new AI users” |
| Super | $100/mo | 15x more usage than Plus; priority model access at peak times | “Recommended for daily agent users” |
| Ultra | $200/mo | 50x more usage than Plus; earliest access to products and models | “Recommended for power users” |
How it's counted: One pool of AI usage shared across Cursor and Grok; each product uses it at a different rate. Third-party models are billed separately at API list price. Resets: Your allowance refreshes weekly. At the limit: Keep going with top-offs; wait for your weekly refresh; move to a bigger plan. Plus includes Grok models only. You can set spending caps for third-party models.
Sources (checked 2026-11-03): Pricing and FAQ (opens in a new tab) · ToolPick review
GitHub Copilot plans and limits
Our pick by use: A few times a week: Free · Every day for work: Pro · All day or with agents: Pro+ · Open in the picker
| Plan | Price | What the plan includes | GitHub Copilot says it's for |
|---|---|---|---|
| Free | $0 | 2,000 code completions a month and an allowance of AI credits; auto model selection only | “Try AI-powered coding assistance at no cost” |
| Pro | $10/mo | 1,500 AI credits a month (1,000 base + 500 flex); unlimited completions | “For individuals who want more flexibility with a selection of models and a monthly allowance of AI credits” |
| Pro+ | $39/mo | 7,000 AI credits a month (3,900 base + 3,100 flex); premium models | “Ideal for AI power users who want access to the most advanced capabilities” |
| Max | $100/mo | 20,000 AI credits a month (10,000 base + 10,000 flex); priority access to premium models | “Ideal for sustained, high-volume AI power users” |
How it's counted: Monthly GitHub AI Credits (1 credit = $0.01); each chat or agent task costs credits by model and tokens. Code completions and next edit suggestions are unlimited on paid plans and use no credits. Resets: Monthly, at 00:00 UTC on the 1st of each calendar month. Unused credits don't carry over. At the limit: Upgrade and pay only the price difference; your new allowance applies right away; set a budget and pay for extra usage at $0.01 per credit; wait for next month's reset. Credits = base credits (match the price) + a flex allotment GitHub can adjust.
Sources (checked 2026-11-03): Plans (opens in a new tab) · Usage-based billing (opens in a new tab) · ToolPick review
Windsurf plans and limits
Our pick by use: A few times a week: Free · Every day for work: Pro · All day or with agents: Max · Open in the picker
| Plan | Price | What the plan includes | Windsurf says it's for |
|---|---|---|---|
| Free | $0 | Light quota to code with agents; unlimited Tab completions and inline edits | — |
| Pro | $20/mo | Increased quotas; all models, including frontier models | — |
| Max | $200/mo | Significantly higher quotas | “For power users who need significantly higher usage quotas” |
How it's counted: A usage quota per plan; each message costs more or less depending on the model, task size and reasoning. Resets: Refreshes automatically every day and every week. At the limit: Buy extra usage, charged at API pricing; wait for the daily or weekly refresh. Windsurf says most users never hit their limits. Smaller models make the quota last longer.
Sources (checked 2026-11-03): Pricing and FAQ (opens in a new tab) · ToolPick review
Tips to make a limit last longer
- Start a new chat for a new topic. Claude and Gemini both count long chats as more usage.
- Use the lighter model for simple jobs. Gemini, Windsurf and Cursor all say the model you pick changes how fast you use your allowance; ChatGPT's Fast mode uses it 2.5x faster.
- Check the meter. Claude: Settings › Usage. Gemini: Settings › Usage Limits. Microsoft: Services & subscriptions.
- Paying for two? If you only hit limits on one tool, upgrade that one and drop the other. Audit what you pay →
Left out
- Perplexity: its pricing page showed a security check today, so we couldn't read its plan limits.
- Microsoft 365 Pro: the plan is named on Microsoft's limits page, but its price isn't on the plan comparison page we read.
FAQ
Does ChatGPT, Claude or Gemini have a fixed number of messages?
Mostly no. Claude and Gemini count usage by how much work a request takes (chat length, files, model), so there's no fixed message count. ChatGPT says everyday text chats are unlimited on every plan, with separate limits for uploads, images, deep research, Thinking and Codex. GitHub Copilot is the clearest: a set number of AI credits a month (1,500 on Pro).
Which AI limits reset every 5 hours?
Claude (a rolling 5-hour session plus a weekly limit), Google Gemini (every 5 hours until a weekly limit) and ChatGPT's Codex and Work on Plus (5 hours plus weekly). Cursor resets weekly, Windsurf daily and weekly, and GitHub Copilot monthly on the 1st (00:00 UTC). Checked 2026-11-03.
What happens when I hit my Claude limit?
Claude says you can wait for it to reset, move to a higher plan, or on paid plans turn on usage credits to keep working at standard API rates. Claude Code uses the same limit as chat.
Is Claude Max or ChatGPT Pro worth it for heavy use?
Both start at $100 a month. Claude Max gives 5x Pro usage every 5 hours and every week. ChatGPT Pro 100 adds Pro reasoning and removes the 5-hour limit in Codex and Work. If you mostly code with agents, compare them with Cursor Super ($100, 15x Plus) and GitHub Copilot Max ($100, 20,000 AI credits).
Can I pay for extra usage instead of upgrading?
Claude (usage credits at API rates on paid plans), GitHub Copilot ($0.01 per AI credit with a budget you set), Cursor (top-offs), Windsurf (extra usage at API pricing) and ChatGPT (credits for Codex, Work and Office add-ins) all say yes. Gemini's limits page doesn't mention a top-up. Mistral and Microsoft don't say.
How did you pick the plan for each use level?
From each company's own wording where it gives one (Claude: Pro “for everyday work”, Max “for all-day work”; Cursor recommends Premium for new AI users and Super for daily agent users; GitHub calls Pro+ ideal for AI power users), otherwise from the allowances each plan states. We don't guess message counts the companies don't publish.
Related: Free plan limits · ChatGPT pricing · Cursor pricing · ChatGPT vs Claude · Cursor vs GitHub Copilot vs Windsurf · Monthly or yearly? · How we pick