What are Grok Imagine’s generation limits?
Updated August 2026
Ask an AI about this guide
Open this article in your assistant with a ready-made question about doing the same job on your own API keys.
xAI is unusually quiet about exact quotas, so most of what is known about Grok Imagine limits comes from consistent user reports rather than published numbers. Here is the mid-2026 picture, with the appropriate uncertainty flagged, and what to do when the limits are the problem.
Reported limits by tier
xAI applies dynamic, fair-use throttling that shifts with load, so all figures below are user-reported ranges, not commitments.
- Consumer quotas vary by subscription, feature and current demand; xAI does not publish one durable generation count for every account.
- Video consumes more capacity than images, so it can reach a limit sooner during the same creative session.
- Reset windows and throttling can change. Treat the quota shown in your current account as the source of truth.
- API usage follows provider billing and rate limits instead of a consumer-app generation allowance.
The quirks that make Grok limits feel unpredictable
Because throttling is dynamic, the same account can get generous capacity on a quiet morning and hit a wall the same evening. There is no quota meter, so the first signal you get is a slowdown or refusal. Video is rationed much more tightly than images - clips cost far more compute - and users consistently report video caps arriving well before image caps.
This fuzziness is a reasonable design for a consumer product riding demand spikes; it is a poor foundation for any workflow with a deadline.
API access is limited by spend, not by a rolling window
Grok image and video generation are available through the xAI API, billed per generation. There, capacity is a function of your budget and standard API rate limits rather than a shifting fair-use algorithm - which is the structural fix if you keep getting throttled mid-project.