Direct answer: no dependable current free-user quota was verified
For Grok’s consumer website and apps, current official materials describe free access as limited, but the materials reviewed do not specify a numeric free-user allowance or exact reset interval. The official FAQ says free Chat and Voice limits reset on their own schedule; it does not give the schedule or a free-user usage-check path. These statements concern grok.com and the Grok apps, and do not establish the limit for Grok on X. (Official Grok FAQ; official plan page)
The dated reports do not fill that gap: an @grok post dated July 19, 2025 says free usage typically resets every two hours, with 20–30 queries per period depending on complexity. A third-party article dated October 25, 2026 says xAI publishes no numeric free-tier count and describes free Chat and Voice limits as resetting on their own schedules. Neither establishes a guaranteed current allowance, and the available evidence does not show whether they concern the same surface or policy. The 2025 post does not define “query” or explain how complexity affects the count. (2025 @grok post; October 2026 report)
Capacity-planning takeaway: Don’t use the dated 20–30 figure as a guaranteed budget. Because the current number and exact reset interval are unverified, this evidence cannot establish whether free access will cover a particular workload.
What happens when you reach the free limit?
The 2025 @grok post says to wait for a reset and says there is no permanent lockout. Current official material reviewed here does not specify what happens when a free user reaches a limit, when access returns for an individual account, or where a free user can check usage or reset status. The official FAQ’s statement about paid features pausing at a weekly limit concerns paid usage; it separately says free Chat and Voice limits reset on their own schedule. That is not a documented free-user check path. (Official Grok FAQ; 2025 @grok post)
Listed consumer-plan prices
The official plan page lists these dollar-denominated monthly amounts and relative usage levels. It says prices exclude applicable taxes or fees and are subject to change. Check the purchase terms shown for your account before subscribing. (xAI pricing page)
| Plan | Listed amount | Listed usage information |
|---|---|---|
| Free | $0/month | Limited usage, including limited Chat requests |
| Plus | $8/month | Basic usage level |
| Premium | $30/month | 4× Plus usage |
| Super | $100/month | 15× Plus usage |
| Ultra | $200/month | 50× Plus usage |
The multipliers are relative comparisons, not message counts or guaranteed allowances. The official FAQ describes paid usage as a weekly pool and distinguishes it from free Chat and Voice limits, which reset on their own schedule. It does not provide a numeric free allowance. (xAI pricing page; official Grok FAQ)
Budget takeaway: Plus is the lowest listed paid amount on the plan page, but the available evidence does not establish a guaranteed message count for it. The 2025 post’s “unlimited access” wording is not enough to confirm an unlimited entitlement or its price.
Consumer access and xAI API costs are separate
Consumer-plan limits are not API pricing or API rate limits. The xAI developer documentation excerpts reviewed describe API token pricing and separate rate limits; they do not make a consumer subscription’s allowance an API allowance. (API pricing; API rate limits)
The supplied API pricing information lists Grok 4.7 at $2 per million input tokens and $6 per million output tokens below the long-context threshold. Using those listed rates, a hypothetical request with 100,000 input tokens and 20,000 output tokens would cost about $0.32: (100,000 ÷ 1,000,000 × $2) + (20,000 ÷ 1,000,000 × $6). This is an arithmetic example, not a measured workload or a full bill estimate; model, context length, token use, and any applicable tool charges can affect cost. Check current API pricing before budgeting.
The supplied rate-limit information says API limits depend on model and team tier and are measured in requests per second and tokens per minute; exceeding a limit can return a 429 response. For production, estimate cost and throughput using your expected model, input and output volume, and request pattern. Without those workload details, it isn’t possible to say whether the API will be cheaper or sufficient for your use case.
Sources
@grok post on free usage limits, July 19, 2025