MiniMax H3.1: Is It Released, and Is It the Same as H3?
MiniMax H3.1 is on the reported second-half roadmap, but has no public API model ID or price. Distinguish it from the available H3 video model and M3 text API.
Published 2026-09-27
MiniMax H3.1 has not been released as a documented public API as of September 27, 2026. It was named alongside M3.1 and M3 Pro in an earlier second-half product pipeline discussion, but there is no final H3.1 model card, accepted API ID, price, context or output specification, or launch date. The H3 and H3 Max names in MiniMax's September 28–October 7 daily check-in promotion do not mean H3.1 has launched.
Do not confuse H3.1 with existing MiniMax H3 video generation. APIMaster's H3 video guide covers a different, already launched product billed by video duration, not the unreleased H3.1 tier. For a callable MiniMax text API now, minimax-m3 is listed from $0.30 input / $1.20 output per 1M tokens across 3 active routes, against its marketplace original $0.60 / $2.40 (about 50% off, checked September 27). View the M3 discount; the live card controls current terms.
| Model ID | Original price (1M input / output) | APIMaster current price | Active routes | Discount | Model card |
|---|---|---|---|---|---|
| H3.1 (API ID unannounced) | Not published | No public route | 0 | — | Not listed |
minimax-m3 (available text API) |
$0.60 / $2.40 | From $0.30 / $1.20 | 3 | ~50% off | View discount |
GPT Trial(GPT-6 Astra, GPT-6 Sol, GPT-5.6, and GPT-5.5 available)
Sign up and claim a $20 GPT trial
Has MiniMax H3.1 launched?
No documented public H3.1 endpoint or price was available in the September 27 check. A place in a roadmap signals planned work, not an API commitment. In particular, do not infer an H3.1 context length, output cap, supported modality, parameter count, token rate or model ID from the name alone.
The September 27 release is M3.1-Flash-Preview, accessible in MiniMax Code for everyday coding. The accompanying Token Plan quota reset and September 28–October 7 double daily free check-in quota mention M3.1-Flash-Preview, H3 and H3 Max; they do not list H3.1. Those plan benefits are not an APIMaster API price. The M3.1 guide explains the released preview and the pending full-model API separately.
H3.1 vs H3 and H3 Max: the naming trap
H3.1 is a future roadmap name. MiniMax H3 is an existing video generation model with an APIMaster guide covering asynchronous text-to-video tasks, 768P and 2K options, and per-second billing. H3 Max is mentioned in MiniMax's check-in benefit. Neither name authorizes us to copy a video price or capability into a supposed H3.1 token price.
| Name | What is actually established | What not to assume |
|---|---|---|
| H3.1 | Named on the second-half roadmap; public API details pending | That it is already live, or that H3's video pricing applies |
| H3 | Existing video product and APIMaster integration | That a video model is the H3.1 public API |
| H3 Max | Named in the limited-time MiniMax check-in offer | That the promotion reveals H3.1's specs or API status |
| M3.1-Flash-Preview | Launched September 27 in MiniMax Code | That it is an H-series release |
Even H3.1's intended modality should wait for provider documentation. A decimal suffix is not enough to infer that it is a drop-in upgrade for the existing H3 video endpoint. Treat a community API ID or pricing chart as a lead for verification, not a migration instruction.
What can developers use while waiting?
For video generation, use the existing MiniMax H3 guide and confirm its live duration-based card. This article deliberately does not place H3's per-second video price in a per-million-token table: the two billing units cannot be compared directly.
For text and coding via API, use minimax-m3. On September 27 its cheapest APIMaster active route listed $0.30 / $1.20 per 1M tokens, with 3 active routes. Its marketplace original was $0.60 / $2.40, approximately 50% off. This is an M3 price, not an H3.1 quote. See the M3 card for current route-specific billing and availability.
To test M3, register, create an API key, and use the listed ID:
curl "https://apimaster.ai/v1/chat/completions" \
-H "Authorization: Bearer $APIMASTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"minimax-m3","messages":[{"role":"user","content":"Summarize the likely cause of this error."}]}'
This is not an H3.1 sample. Keep the later switch behind a configuration value until MiniMax publishes a callable H3.1 interface.
What will make an H3.1 API announcement actionable?
Look for an official model card that states the release stage, exact request ID, endpoint, modality, input format, output format and rate limits. Then confirm official pricing in the appropriate unit (tokens, seconds or something else), caching rules, context and output limits, and compatibility with your existing client. Finally, make a low-volume request and check its response and invoice against the published documentation.
If the product turns out to belong to a different family from existing H3 video generation, the benchmark should reflect that. A single ranking against unrelated text or video models would be misleading. The M3 Pro tracking page similarly keeps reported architecture claims separate from a shipped API.
FAQ
Is MiniMax H3.1 available now?
No public H3.1 API identifier, price or model card was published by the September 27 check. H3 and H3 Max names in a check-in offer do not change H3.1's release status.
Is H3.1 the same model as MiniMax H3?
No equivalence has been announced. Existing H3 is an available video-generation product with duration-based billing; H3.1 is a distinct roadmap name without published public API specifications.
What will H3.1 cost?
There is no official H3.1 API price yet. M3's $0.30 / $1.20 per 1M starting rate is a price for an existing text API, not a forecast for H3.1.
Does the double free-quota promotion include H3.1?
The listed models are M3.1-Flash-Preview, H3 and H3 Max for September 28–October 7. H3.1 is not named; confirm live MiniMax plan terms before relying on the promotion.
Sources
- MiniMax — official product announcements and plan details.
- MiniMax H3 release notes — existing video model, not an H3.1 release announcement.
- APIMaster H3 video guide — current video integration and billing unit.
- APIMaster M3 live card — text API route count and pricing checked September 27, 2026.
For a working MiniMax text API, register with APIMaster, create a key and call minimax-m3 at https://apimaster.ai/v1. Revisit H3.1 only after an official model card identifies its actual endpoint and price.