GPT Image 2 Rate Limits & Usage Caps
API rate limits, per-minute and per-day caps, and usage tier requirements for GPT Image 2 as documented at the April 2026 launch.
TL;DR
GPT Image 2 rate limits follow OpenAI's standard usage-tier system (Tier 1–5) and are denominated in images per minute (IPM) or requests per minute (RPM). Exact IPM/RPM values differ by tier and are subject to change by OpenAI. The figures below reflect the April 2026 launch documentation; check the OpenAI platform rate-limits page for current values.
Key facts
- →Rate limits are tier-dependent — Tier 1 (new accounts) have significantly lower limits than Tier 4–5 (high-spend accounts)
- →Limits are enforced per API key and reset on a rolling per-minute basis
- →Image generation and image editing share the same quota pool for your account tier
- →Batch requests are not documented for the Images API as of April 2026
- →Hitting a rate limit returns HTTP 429 — implement exponential backoff in production
- →Usage caps (spend limits) can be set in the OpenAI platform dashboard under 'Limits'
- →Always fetch current limits from platform.openai.com/account/limits rather than relying on third-party sources
FAQ
What is the images-per-minute limit for GPT Image 2?
OpenAI's exact IPM limits per tier are published on the platform.openai.com/docs/guides/rate-limits page and change as the service scales. Check there for the current figures rather than relying on figures published here, which could become outdated.
How do I request a rate limit increase?
Log in to platform.openai.com and navigate to the rate limit increase request form. Increases are evaluated by OpenAI based on usage history and business justification.
Do failed requests count against my rate limit?
Requests that are rejected at the API gateway (before reaching generation) typically do not count against image quotas, but the request itself may count toward RPM. Check the OpenAI rate-limits documentation for the definitive answer.
Is there a daily cap separate from per-minute limits?
OpenAI enforces per-minute rate limits and monthly spend caps. A separate daily image cap has not been documented for GPT Image 2. Your spend limit (configurable in the dashboard) is the primary absolute ceiling on usage.