[Aug 24–Sep 6 | 14-Day Free] MiniMax Free API Guide: Unlimited M3/M2.7 on GMI Cloud + OpenRouter

Updated: 2026-08-26 | Platform: https://platform.minimax.io | Official: https://www.minimax.io

🔥 Limited Free · Aug 24–Sep 6 (14 days · deadline announced): MiniMax 14-day unlimited FREE on GMI Cloud + OpenRouter — M3 / M2.7 / Speech 2.8 / Music 3.0 at $0 (official deadline Sep 6, 2026, see §3.5)

MiniMax is one of the few labs building frontier models across text, speech, video, image and music (M3/M2.7, Hailuo H3, Speech 2.8, Music 3.0, 200M+ users). This guide summarizes the latest official pricing into how to use MiniMax at the lowest cost — or near zero cost.

1. TL;DR: No permanent free tier, but 4 low-cost / near-free paths

Path	For	Cost	Key point
M3 permanent 50% off PAYG	Variable usage	from $0.30/M input	Long-term discount, no coupon
Token Plan subscription	Heavy Agent/coding/multimodal	$22/$55/$132 per mo	One plan covers text+image+speech, 5h rolling + weekly quota
Credits prepaid packs	Flexible pay-as-you-go	1,000 Credits = $1	$5/$25/$100, 365-day validity, auto-cover overflow
Limited promos & referral	New users & inviters	10% off first order + 10% voucher back	Co-builder program Dec 26, 2025 - Aug 31, 2026

Site check: no MiniMax article among 54 free-API posts on apishare.cc — this is the first.

2. What you get

LLM: MiniMax-M3 (512K context, 1M+ extended), M2.7/M2.5/M2.1, Anthropic/OpenAI/AI SDK compatible, Prompt Caching, Tool Use, Interleaved Thinking.
Video: Hailuo H3 (2K, first/last-frame, subject-reference, T2V/I2V)
Speech: Speech 2.8-turbo/hd, sync/async, WebSocket streaming, voice clone & voice design
Image: T2I / I2I
Music: Music 3.0 + Lyrics generation

A few specials (H3, voice design, rapid clone) are PAYG/Credits only, not included in Token Plan quota.

3. Pricing breakdown (latest official)

3.1 Pay-as-you-go (API Key, balance deducted in real time)

M3 Standard (after permanent 50% off)

≤512K input: $0.30/M in, $1.20/M out, $0.06/M cache read
512K input: $0.60/M in, $2.40/M out, $0.12/M cache read
Priority tier (service_tier=priority): 1.5x standard

M2.7: $0.30/M in, $1.20/M out, $0.06/M cache read, $0.375/M cache write; highspeed is 2x

Audio: T2A turbo $60/M chars, hd $100/M chars; rapid clone $1.5/voice, voice design $3/voice

3.2 Token Plan (Subscription Key, usage bar in console)

Tier	Price	Best for	Agents	Windows
Plus	$22/mo	Personal projects	3-4	5h rolling + weekly
Max	$55/mo	Daily Agent coding + multimodal	4-5	same
Ultra	$132/mo	Heavy Agent long sessions	6-7	same

Quota covers text/image/speech (excl. H3 etc.). Overflow auto-covered by Credits if available. Monthly quota does not roll over.

3.3 Credits

1,000 Credits = $1, deducted at PAYG list price
Packs: $5=5,000, $25=25,000, $100=100,000, 365-day validity
Priority: Token Plan quota first, then Credits

3.4 Promos

10% off first order via referral link (all tiers/upgrades, multiple purchases during event)
10% voucher back to inviter (90-day, API-fee only)
$2 Starter ended Jan 15, 2026 — watch for reruns

3.5 Latest Free Access (Verified 2026-08-26 · Aug 24–Sep 6 · 14-Day Pinned · Deadline Announced)

Pinned · 14-Day FREE Aug 24–Sep 6 (deadline announced): Verified via Brave search + official X posts — @MiniMax_AI Aug 24: "From Aug 24–Sep 6, get unlimited access to MiniMax M3 and M2.7 on @gmi_cloud—plus Speech 2.8 and Music 3.0" and @gmi_cloud: "14 days FREE on GMI Cloud from 8/24 to 9/6, access via both GMI API key or @OpenRouter". Window is Aug 24–Sep 6, 2026 (14 days), also mirrored on OpenRouter minimax/minimax-m3:free & minimax/minimax-m2.7:free (pricing: {"prompt":"0","completion":"0"}). Use it for long-context + tool-use + caching + speech/music E2E, then fall back to M3 50% off or Token Plan.

Added after web-wide check on 2026-08-26. Always confirm in console — promos change without notice.

A. Official API new-user trial ¥15 (~$2)

Sources: yangmao.ai free-API tracker (snapshot 2026-06-24) and Apidog How to Use MiniMax M2-7 for Free (2026) both confirm ¥15 trial credits on signup.
Covers: MiniMax-M2.7 chat, TTS, etc.; OpenAI-compatible endpoint https://api.minimaxi.chat/v1/chat/completions (also https://api.minimax.io/v1).
How: Sign up at https://www.minimaxi.com / https://platform.minimax.io → check balance → run the minimal OpenAI SDK request below.
Caveat: Quota/RPM/expiry may change; verify in dashboard before production.

B. Hailuo H3 — four free doors (all capped at 768p)

Sources: PixMind MiniMax H3 Free (2026-08-01) and AtlasCloud Is MiniMax H3 Free? Four Free Doors (2026-08-04), tested.
Doors: 1) Hailuo official free tier 2) MiniMax Hub trial 3) Open-weights self-host (42.5GB+) 4) Aggregator trials (OpenRouter / Hugging Face Spaces / PixMind & AtlasCloud starter credits).
Limits: Free = 768p max, watermark on some; 2K watermark-free is paid: $0.13/sec at 2K, $0.09/sec at 768p (AtlasCloud lists $0.14/sec on 2026-08-04; 15s 2K ≈ $1.95), linear billing, no bulk discount.
Advice: Use free to lock the first frame and test the workflow; ship paid for watermark-free 2K.

C. Aggregator free lanes (no-card quick test)

OpenRouter: Unified OpenAI format for minimax/minimax-m2.7, great for multi-model compare & fallback.
Hugging Face Spaces: Community demos, free but queued/limited.
PixMind / AtlasCloud: Starter credits on signup, good for no-card H3 validation — check licensing & reliability for commercial use.

Tip: Burn the ¥15 on M2.7 coding/tool-use & long-context first, then use H3 free 768p for storyboards, then pay $0.13/sec for a single 2K deliverable — cheapest way to compare.

4. Zero-cost quickstart in 3 steps

1) Get a key

2) Env (Anthropic recommended)

export ANTHROPIC_BASE_URL=https://api.minimax.io/anthropic
export ANTHROPIC_API_KEY=YOUR_KEY
# or OpenAI:
export OPENAI_BASE_URL=https://api.minimax.io/v1
export OPENAI_API_KEY=YOUR_KEY

3) Call

import anthropic
client = anthropic.Anthropic()
resp = client.messages.create(model="MiniMax-M3", max_tokens=1024, messages=[{"role":"user","content":"Introduce MiniMax M3 in one sentence"}])
print(resp.content[0].text)

Cache repeated context to cut read cost to $0.06/M.

5. Which to pick?

<50M tokens/mo & spiky: PAYG with M3 50% off
Daily Agent coding/multimodal: Max/Ultra subscription
No monthly commitment: Credits packs (team-shareable)
Teams: Token Plan for Teams with seat & shared Credits pool

6. On apishare.cc

Apishare aggregates free inference channels. MiniMax has no permanent free tier officially, but you can search apishare.cc for community-contributed free proxies or combine official M3 discount + Token Plan for dev-low-cost + prod-elastic.

7. FAQ

Free tier? No permanent free tier. Closest is M3 50% off, $22 Plus covering full stack, plus referral 10% off/voucher.

Keys interchangeable? No. API Key = PAYG, Subscription Key = Token Plan/Credits.

Over quota? Use Credits, upgrade, switch to PAYG key, or wait for window reset.

Cursor/Claude Code/OpenClaw? Follow https://platform.minimax.io/docs/token-plan/claude-code etc. — just set Subscription Key + Base URL.

Sources: MiniMax Pricing Overview, PAYG, Token Plan, Token Plan Overview, Prerequisites, Referral Program (fetched 2026-08-26). Always confirm on official site.

Need a shared MiniMax proxy or team plan pooling on apishare.cc? Contact verify@samai.cc.

🚀 Get Started: One-Click Free API Access

Want to call all the free models above with a single API key, no need to sign up for each provider? Apishare.cc provides a unified API Key — one key, 100+ models, free models at zero cost.

👉 Register on Apishare.cc → Get your unified API Key

📊 Want to see more free model rankings? Check out the Sep 2026 Free LLM API Rankings →

About the Free API Aggregator

The models covered in this guide are all served through the APIShare free API aggregator, which gives you one key for the whole catalog.

Full model catalog: APIShare free API directory
Sign up for a free trial key: Register and claim your API key

[Aug 24–Sep 6 | 14-Day Free] MiniMax Free API Guide: Unlimited M3/M2.7 on GMI Cloud + OpenRouter | M3 50% Off + Token Plan