AI usage limits: ChatGPT, Claude, Gemini and Grok

Prices checked Nov 4, 2026 · Updated Nov 4, 2026

How each vendor meters usage

ChatGPT: per-model allowances, 5-hour window plus weekly caps

• Chat text is effectively unlimited on Free and Go (abuse guardrails apply).
• For Work and Codex on Plus and Business Standard, usage is a 5-hour window measured per model, with weekly limits on top.
• Work and Codex share one allowance.
• Estimates, not fixed limits.

• Pro plans: no 5-hour limit. Weekly GPT-6 Pro message cap in Chat. Pro 200 > Pro 100 in Work/Codex allowance; Pro 500 is highest.
• Fast mode uses your included allowance at 2.5x the Standard rate. Ultrafast (GPT-6 Astra, GPT-6.1 Sol; Pro 500 only) uses 8x. With purchased credits: Fast 2x, Ultrafast 6x.
• Image generation drains included limits about 3-5x faster than a similar turn without images.
• Credits can be bought on Plus, Pro and Business to continue Work, Codex and the Word/Excel/PowerPoint extensions after the allowance runs out. Credits do not extend every ChatGPT feature.
• Desktop voice in Work/Codex costs $0.05 per minute of your Codex budget (1.25 credits per minute on credit-based workspaces).
• Business Standard: 15 GPT-6 Pro messages per month. Business Premium: 50 per week (shared with GPT-5.6 Sol Pro). Premium seats have no 5-hour limit.

Source: ChatGPT Work / Codex pricing and usage, ChatGPT Business models and limits, ChatGPT Pro tiers (Help Center).

Claude: one shared pool, 5-hour session window plus weekly limit

• Your usage across Claude on web, desktop, mobile and Claude Code draws from the same pool.
• Most accounts have a limit on a rolling 5-hour session plus a weekly limit.
• Anthropic gives no fixed message count: it depends on model, conversation length and file size.

• Free: limited usage per 5 hours and per week.
• Pro: at least 5x Free per 5-hour session (1x baseline for the multipliers below).
• Max: 5x Pro every 5 hours and every week.
• Ultra: 10x Pro weekly, no 5-hour limit for verified accounts.
• Fable (top model): Pro pays with usage credits only; Max, Ultra and Team Premium seats can use it for up to 50% of the weekly limit.
• Team Standard seat: 1.25x Pro per-session usage; Premium seat: 6.25x Pro. Limits are per member (no shared drain).
• On paid plans you can turn on usage credits to continue at standard API rates when you hit a limit.
• Anthropic may apply other limits (weekly and monthly caps, model and feature limits) at its discretion.

Source: Claude pricing page, What is the Team plan? (Claude Help Center).

Gemini: compute-based limits, refreshed every 5 hours plus a weekly limit

• Gemini Apps limits are compute-based: they depend on prompt complexity, the model and features you use, and chat length.
• Your limit refreshes every 5 hours until you reach your weekly limit.
• Google publishes multiples, not absolute numbers.

• Free = standard limits. AI Plus = 2x. AI Pro = 4x. AI Ultra = 5x or 20x AI Pro (that is 20x or 80x Free).
• Relative cost per prompt: simple chat is very low; long chats and large files are medium; Deep Research and video generation are high (video uses the most).
• With a Google AI plan, when you hit the limit you can keep chatting with Flash-Lite, or wait for the 5-hour or weekly refresh.
• Without a plan, heavy features such as Deep Research may be unavailable at peak demand.
• Context window: 32K tokens on Free, 128K on AI Plus, 1M on AI Pro and AI Ultra.
• Google says limits may change without notice, including for capacity reasons.

Source: Gemini Apps limits and feature availability (Google Help), Google AI plans (gemini.google).

Grok: one weekly pool shared with Cursor

• Each paid seat gets a weekly pool for first-party usage, shared across Grok (Chat, Imagine, Voice, Bot) and Cursor.
• It resets weekly and does not roll over.
• Different activities drain different amounts.

• Plus = Base (amount not published). Premium = 4x. Super = 15x. Ultra = 50x.
• Grok Bot draws from the same pool, not a separate one.
• Third-party models (Claude, GPT, etc.) in Cursor are never included. Billed at API list price from the first use, from prepaid credit, Premium and up only.
• When the pool runs out: wait for the weekly reset or buy top-ups (Additional Usage credits). Monthly spending caps and alerts are available.
• Teams: each seat has its own pool (not shared across seats). The team shares one prepaid balance for Additional Usage.
• Teams add a Cursor Token Rate: $0.25 per million tokens on third-party model requests, on top of the API price. Individual plans do not pay it.
• Exact message, image and video counts are not published.

Source: X Subscription usage limits (xAI docs), x.ai pricing page, The X Subscription pricing (xAI docs).

Usage by plan

Find a plan that fits your usage · See every plan on one page

Sources

ChatGPT Work / Codex pricing and usage · ChatGPT Business models and limits · ChatGPT Pro tiers (Help Center) · Claude pricing page · What is the Team plan? (Claude Help Center) · Gemini Apps limits and feature availability (Google Help) · Google AI plans (gemini.google) · X Subscription usage limits (xAI docs) · x.ai pricing page · The X Subscription pricing (xAI docs)