CheaperGen
MiniMaxLive status

MiniMax H3 Fast API status

Operational

Is MiniMax H3 Fast down right now? No. MiniMax H3 Fast is up and completing generations normally.

Last check
99.91% uptime over 30 days

Uptime

MiniMax H3 Fast uptime

Each bar is a day (or an hour) of checks every five minutes and finished generations; only failures on our side count. Hover or tap a bar for its details. All models' status

24 hours
99.90%
No incidents
7 days
99.91%
No incidents
30 days
99.91%
No incidents
90 days
99.91%
No incidents

Last 30 daysLast 90 days

  • Uptime
  • ≥ 99%
  • 95–99%
  • < 95%
30 days ago90 days ago99.91% uptime99.91% uptimeToday

Last 24 hours

24 hours ago99.90% uptimeNow

Speed

Generation time

Median time from submitting a request to a finished output, per task, over the generations that succeeded in the last 7 days. Across all tasks in the last 24 hours: 14s. Resolution, length and queue all change it.

History

Recent incidents

An incident is a stretch of hours in which a check found the model unusable, or at least a quarter of its generations (two or more) failed on our side. Times are in UTC.

No incidents in the last 90 days.

If it's down

If MiniMax H3 Fast is down

CheaperGen runs official models below list price through lab partnerships, our own GPU inference for open-source models and automatic routing. The routing is what keeps MiniMax H3 Fast running when capacity fails.

Automatic routing

We run MiniMax H3 Fast across several GPU inference pools and check them every five minutes. When one slows down or fails, new requests switch to a healthy one automatically: same endpoint, same request. If nothing can take a request, it's declined up front and nothing is charged.

Failed generations are refunded

If a generation fails, the price held when you submitted goes back to your balance in full, automatically. The error says why and whether a retry can help. Error handling

A disclosed emergency price

When only pricier capacity is left, a request can run at the emergency price: never above the official MiniMax H3 API price, and shown before it runs. On the API, allow_fallback_price: false keeps you at the standard price. Emergency price

More from MiniMax

Similar video models to try

FAQ

MiniMax H3 Fast status questions

Every model's live state is on the status page.

No. MiniMax H3 Fast is up and completing generations normally. Last checked Oct 25, 2026, 20:20 UTC; uptime over the last 30 days: 99.91%. This page updates every five minutes.

Every five minutes we check whether MiniMax H3 Fast can be served, and we count every generation that finishes. Uptime is the share of those observations that found it working: 1 − (checks that found it unusable + generations that failed on our side) ÷ (all checks + all finished generations). Failures caused by the request itself, like a blocked prompt or an unreadable file, don't count, and degraded performance or the emergency price still count as up. A day with nothing recorded counts as fully available. Bars are green from 99%, amber from 95% and red below.

No. CheaperGen is independent of MiniMax, and this page doesn't monitor MiniMax's own console or API. It shows MiniMax H3 Fast as served by CheaperGen on optimized GPU inference, measured by our own health checks and the generations that run here: whether MiniMax H3 Fast works on CheaperGen right now.

It's refunded automatically: the price held when you submitted goes back to your balance in full, and the generation's error says why and whether retrying can help. When one inference pool fails, new requests switch to a healthy one automatically. If nothing can take a request, it's declined before anything is charged.

Run MiniMax H3 Fast now

Generate in the Studio, or call MiniMax H3 Fast from your code with one API key. Failed generations are refunded automatically.

Open in Studio API reference
curl -X POST https://cheapergen.com/api/v1/run/minimax/h3-fast/text-to-video \
  -H "Authorization: Bearer $CHEAPERGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "…"}'