MiniMax M3
New$0.30
$1.20
- Positioning
- Balanced. The vendor doesn't give one, so it's placed by price among text models.
- Price
- Level 2 of 4 among text models ($ is the cheapest), by output price per 1M tokens: $1 to under $5.
- Vendor-stated speed
- Not shown: the vendor doesn't compare this model's speed.
Three separate signals, not a quality score.
MiniMax's long-context workhorse — two price bands either side of 512K input. MiniMax M3 is a text model on ATP Token, billed per 1M input and output tokens in input-length tiers, with a lower cached-input rate.
Pricing
| Input length | Input | Cache read | Output |
|---|---|---|---|
| ≤512K | $0.30 | $0.06 | $1.20 |
| >512K | $0.60 | $0.12 | $2.40 |
Each request's total input (cached tokens included) picks the band. That band's price covers the whole request, output too.
Billed in two input bands: a request whose input exceeds 512K tokens bills at double the rate across input, output and cache read. The band is decided by total input tokens, cache writes included. Passthrough of the model's official list price — no ATP markup.
Cost per 1,000 typical requests
| Workload | USD / 1,000 |
|---|---|
| Short chat turn800 / 400 | $0.72 |
| RAG answer4,000 / 500 | $1.80 |
| Agent step with tools20,000 / 1,500 | $7.80 |
Same controls on every model: project keys, per-project model access, credit budgets, and a log of every request.Set up your organization
Cost estimator
- Input$3.00
- Output$2.40
$5.40 USD / month · 540 credits / month
Quickstart
from openai import OpenAI
client = OpenAI(
base_url="https://api.atptoken.ai/v1",
api_key="atp-•••"
)
r = client.chat.completions.create(
model="minimax-m3",
messages=[{"role":"user","content":"hi"}],
)Compare with similar models
- Qwen 3.7 PlusQwen (Alibaba)Positioning: Balanced. Price level 2 of 4 among text models.$0.40 / $1.60In / out · per 1M tokens
- DeepSeek V3.2DeepSeekPositioning: Balanced. Price level 2 of 4 among text models.$0.57 / $1.71In / out · per 1M tokens
- GLM-5Z.aiPositioning: Balanced. Price level 2 of 4 among text models.$1.00 / $3.20In / out · per 1M tokens
FAQ
- Can MiniMax M3 use tools, read images, or think before answering?
- MiniMax M3 supports tool calling, vision input and extended thinking, but not audio.
- How does MiniMax M3's pricing by input length work?
- You pay per 1M input tokens and per 1M output tokens. The price steps with each request's total input length (≤512K / >512K). The band covers the whole request, output too. Cached input costs 20% of the normal input price. You only pay when a request succeeds. $1 = 100 credits. See pricing
- How do I call MiniMax M3 on ATP?
- Create a project API key and point any OpenAI-compatible SDK at https://api.atptoken.ai/v1. Set model to "minimax-m3" (the API id, not the display name). Make sure the model is turned on for your project.
Start with $5. Credits never expire.
Create a project key, turn this model on, and pay only for what you use.