Gemini shows You've reached your limit, Please try again later, or This feature isn't available right now — don’t assume an outage. As of June 2026, this is almost always a usage-limit or tier issue, not downtime. The fastest fix for most people: if you’re on a paid plan you’ve usually already been auto-dropped to Gemini Flash-Lite, so just keep going; if you’re on the free plan, switch to a lighter model yourself and retry, or wait for the next 5-hour refresh.
The catch: in early 2026 Google replaced the old fixed “X messages per day” caps with compute-based usage limits. Your cap is no longer a message count — it’s an internal budget that drains faster when you use Gemini 3.1 Pro, Deep Research, image/video generation, Deep Think, or very long chats, and refreshes every 5 hours up to a weekly ceiling. So “I only sent 12 messages and I’m already capped” is now expected behavior if those messages were heavy. Figure out which limit you hit first, then decide: switch model, wait, or upgrade.
Which limit did you hit? (decision aid)
| What you see | Most likely cause | Fastest action |
|---|---|---|
You've reached your limit + offered Flash-Lite | Compute budget for Pro drained this 5-hour window | Continue on Flash-Lite, or wait for refresh |
Please try again later that clears in ~1 min | Backend throttle at peak hours | Resend in 1-2 minutes |
| Worked seconds ago, then blocked after rapid sends | Per-minute burst rate limit | Wait 1-5 minutes |
This feature isn't available on Deep Think / video | Feature is tier-gated (Plus/Pro/Ultra) | Upgrade or use a different feature |
| Personal account works, work account doesn’t | Workspace admin set an org cap | Contact your IT admin |
| Capped on day one of a paid plan despite low use | Subscription not recognized / billing flag | Verify plan is Active (see Step 3) |
Common causes
By frequency:
1. Compute budget for the heavy model drained (most common)
Limits are now compute-based, not a fixed message count. Each prompt draws from a budget that depletes faster for Gemini 3.1 Pro, Deep Research, image/video generation, Deep Think, and long conversations. Google does not publish exact numbers; instead it scales the budget by tier (as of June 2026):
| Plan | Price (USD/mo) | Usage budget | When capped |
|---|---|---|---|
| No plan (Free) | $0 | Standard | Wait for refresh, or switch to Flash/Flash-Lite |
| Google AI Plus | $4.99 | 2x standard | Falls back to Flash-Lite |
| Google AI Pro | $19.99 | 4x standard | Falls back to Flash-Lite |
| Google AI Ultra | $99.99 | up to 20x AI Pro | Falls back to Flash-Lite |
Note: “Gemini Advanced” / “Google One AI Premium” was renamed Google AI Pro in early 2026. If an old article or screenshot says “Advanced,” that is now Google AI Pro.
Limits refresh every 5 hours, not at a fixed daily clock, up to a separate weekly ceiling. So a cap can clear in a few hours rather than at midnight.
How to tell: the error mentions “limit” and (on paid plans) offers to continue with Flash-Lite.
2. The feature is tier-gated
Some capabilities require a specific paid plan (June 2026):
- Deep Research — any paid plan (Plus and up)
- Video generation (Veo) — Plus and up
- Image generation / “Redo with Nano Banana Pro” — Plus and up
- Deep Think (extended reasoning) — Ultra only
- Gemini Spark / Daily Brief — Ultra / Plus respectively
- 1M-token context window — included with AI Pro and Ultra
How to tell: the message says the feature isn’t available or prompts you to upgrade, rather than mentioning a limit.
3. Backend throttling at peak hours
During peak load (roughly US-West morning and US-East evening) the backend may briefly throttle requests.
How to tell: Please try again later with no specific limit wording, and it clears within about a minute.
4. Per-minute burst rate limit
Sending ~10+ messages in 30 seconds can trip a minute-level limit even when your daily budget is fine.
How to tell: it worked seconds ago and you’ve just sent many messages in a row.
5. Workspace admin set an org-wide cap
A Workspace (work/school) account can have a domain-wide limit set by IT.
How to tell: your personal account works but your work account doesn’t.
6. Paid plan not recognized / account flagged
Two distinct cases: (a) a billing or provisioning hiccup leaves your paid plan stuck on free-tier limits; (b) heavy automation or repeated safety-filter triggers trip abuse detection.
How to tell: you’re capped immediately on a fresh paid plan with light use (case a), or you’ve been running scripts/bulk prompts (case b).
Shortest path to fix
Step 1: Drop to a lighter model
Model picker (top of the chat) → Gemini Flash, or Flash-Lite
Resend the promptOn paid plans you’re usually offered Flash-Lite automatically once the Pro budget is drained — just accept it and continue. Flash and Flash-Lite draw far less compute, so they last much longer; for everyday tasks the quality gap is small. Save Gemini 3.1 Pro for genuinely hard reasoning.
Step 2: Wait for the refresh
| Error type | Typical wait |
|---|---|
| Per-minute burst limit | 1-5 minutes |
| Compute budget exhausted | Up to the next 5-hour refresh (weekly ceiling permitting) |
| Backend throttle | 1-15 minutes |
Please try again later with no limit wording is usually the per-minute or backend case — retry shortly. A hard “limit reached” follows the 5-hour refresh cycle.
Step 3: Verify your plan is actually Active (if you pay but got capped)
If you’re on a paid plan but hit free-tier limits, confirm billing is healthy before anything else:
- Open myaccount.google.com/subscriptions (or the Google One app).
- Confirm Google AI Pro (or Plus/Ultra) is listed and shows Active — not “payment issue” or “expired.”
- Make sure you’re signed into Gemini with the same Google account that holds the subscription. A second signed-in account is the single most common cause of “I pay but I’m still limited.”
- If it’s Active and you’re on the right account but still capped, report it via the in-app feedback so engineering can check your account.
Step 4: Upgrade your plan
If you legitimately need more headroom, upgrading multiplies your compute budget:
- Google AI Pro — $19.99/mo: 4x the standard budget, 1M-token context, Deep Research, Veo video, Nano Banana Pro images.
- Google AI Ultra — $99.99/mo: up to 20x AI Pro, plus Deep Think and Gemini Spark.
Upgrade at one.google.com/explore-plan or from Settings → Upgrade inside the Gemini app.
Step 5: Use the Gemini API (separate quota)
The API has its own quota, independent of the Gemini app:
from google import genai
client = genai.Client(api_key="YOUR_API_KEY")
response = client.models.generate_content(
model="gemini-3.1-pro",
contents="..."
)
print(response.text)- Free tier: limited RPM / RPD per model; daily request quotas reset at midnight Pacific time.
- Paid tier: token-billed and effectively unlimited for normal use (Gemini 3.1 Pro is
$2in /$12out per 1M tokens as of June 2026).
Your exact RPM/RPD show in AI Studio → rate limits. Note: a paid Gemini-app subscription does not automatically lift your API tier — to get paid-tier API limits you must enable billing on the API key’s Cloud project in AI Studio.
Step 6: Use AI Studio (free, much wider limits)
Open aistudio.google.com
Pick the model, paste your promptAI Studio runs on the API quota system, which is far more generous than the consumer app for free users — good for long-context work, document processing, and prototyping. (Image generation there is also far less restricted than the app’s per-day cap.)
Step 7: Spread heavy work across accounts (only where compliant)
If you have separate personal, work, and student Google accounts with their own budgets, you can split heavy work across them — but only where each account’s terms of service allow it. Do not use this to evade limits in ways the ToS prohibits.
Step 8: Workspace — ask IT to raise the cap
If your work account is the one being limited:
- Contact your IT admin.
- Admin opens Admin Console → Apps → Additional Google services → Gemini app for Workspace.
- Raise the per-user usage allowance for your org or OU.
Step 9: Abuse-flagged — pause, wait, escalate
If automation or repeated safety triggers flagged the account:
- Stop all automation and bulk prompting immediately.
- Wait 24 hours.
- Still blocked → file a ticket at support.google.com/gemini with timestamps.
How to confirm it’s fixed
- The blocked action (a Gemini 3.1 Pro prompt, a Deep Research run, an image) completes without the limit banner.
- The model picker shows you back on your intended model instead of being forced to Flash-Lite.
- If you upgraded: myaccount.google.com/subscriptions shows the new plan as Active, and the in-app usage banner reflects the higher budget.
Prevention
- Front-load heavy Pro/Deep Research work right after a 5-hour refresh, and watch the weekly ceiling.
- Reserve Gemini 3.1 Pro for hard reasoning; use Flash or Flash-Lite for everyday chat to preserve budget.
- High-volume or scripted workloads belong on the API or AI Studio, not the consumer app — the app’s abuse detection flags automation.
- Keep aistudio.google.com bookmarked as a fallback when the app throttles.
- Stay signed into Gemini with the account that actually holds your subscription.
FAQ
Why am I capped after only a few messages? Limits are compute-based now, not a message count. A handful of Gemini 3.1 Pro, Deep Research, image, or Deep Think requests can drain more budget than dozens of simple Flash chats. This is expected as of 2026.
When does my Gemini limit reset? The consumer app refreshes your usage budget every 5 hours, up to a weekly ceiling — not at a fixed midnight. The separate Gemini API resets requests-per-day quotas at midnight Pacific time.
I pay for Google AI Pro but still get limited — why? Most often you’re signed into Gemini with a different Google account than the one holding the subscription, or the plan shows a billing issue. Check myaccount.google.com/subscriptions and confirm the plan is Active on the account you’re using.
Is “Gemini Advanced” the same as Google AI Pro? Yes. Google renamed “Gemini Advanced” / “Google One AI Premium” to Google AI Pro in early 2026. Same paid tier, new name, now including extras like NotebookLM and Flow.
How do I get unlimited Gemini without paying? There’s no truly unlimited free option, but the Gemini API free tier and AI Studio give far more headroom than the consumer app for free users — switch there for long or high-volume work.
Will switching to Flash-Lite ruin my answers? For everyday tasks, no. Flash-Lite is fast and capable; the gap shows mainly on hard multi-step reasoning, where you’d want Gemini 3.1 Pro. Use the right model for the job rather than burning Pro budget on simple prompts.
Related
Tags: #Gemini #Debug #Troubleshooting