Last verified
M3 FRONTIER1M CONTEXTMULTIMODALPROMPT CACHINGSTANDARD + PRIORITY

MiniMax M3 API Pricing

MiniMax M3 is MiniMax's frontier model for coding, agents, multimodal input, and long-context tasks. The official pricing page lists the standard ≤512K row at $0.30/M input, $1.20/M output, and $0.06/M cached input; >512K and Priority rows cost more. Since October 19, 2026 it sits next to the cheaper M3.1 Flash, the only other model left on MiniMax's price list.

Input - per 1M tokens
$0.30/M
Standard ≤512K input row base tier
Output - per 1M tokens
$1.20/M
Standard ≤512K input row 4x input
Cached input - prompt cache read
$0.06/M
Cache read standard row 80% off
Effective - agentic blend
$0.191/M
92/8 split - 82% cache
§ 01 / TERMINAL

Run the numbers.

Live calculator pre-loaded with MiniMax M3 standard ≤512K rates. MiniMax also lists long-context >512K and Priority service-tier rows; use those for very long or high-priority traffic.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

CODING AGENT

Repo repair task

$0.00859/task
60k in / 2k out~11,638 units/$100
OFFICE

Analyst memo draft

$0.00773/memo
40k in / 3k out~12,939 units/$100
SEARCH

Web research brief

$0.00618/brief
25k in / 3k out~16,181 units/$100
CHATBOT

Product support turn

$0.00182/turn
6k in / 1k out~54,969 units/$100
§ 03 / TOKENIZER

Paste text. See tokens. See cost.

What's a token →
Calibrated · measured on the vendor's tokenizer · 2026-06-10 Auto-counts as you type

Counts use a chars-per-token calibration measured on the vendor's own published tokenizer (MiniMaxAI/MiniMax-M2.1, 2026-06-10). English prose is typically within a few percent; code and non-Latin scripts tokenize heavier. For billing-exact counts use the vendor's count-tokens API.

Characters 443
Words 70
Tokens (estimated) 83 tokens
Cost as input · uncached $0.00002 USD
Cost as output · uncached $0.0001 USD
Cost as cached input $0.000005 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
MiniMax M3 Current $0.30 cache $0.06 $1.20 $0.191 agentic 92/8 1M MiniMax frontier coding and agent loops
MiniMax M3.1 Flash $0.15 cache $0.015 $0.60 $0.0842 −2.3× 1M The cheaper, faster sibling with a flat 1M rate
DeepSeek V4 Pro $1.32 cache $0.044 $3.96 $0.569 +3.0× 1M Low-cost reasoning workloads
Gemini 3.8 Flash $0.75 cache $0.075 $3.75 $0.481 +2.5× · promo 1M Global multimodal workloads
GPT-5.4 mini $0.75 cache $0.075 $4.50 $0.541 +2.8× 400K OpenAI subagent workloads

Frequently asked.

Practical MiniMax M3 pricing questions, with live MiniMax list rates separated from workload assumptions.

Q · 01 What is the standard MiniMax M3 API price? +
MiniMax's official pay-as-you-go pricing page lists MiniMax-M3 at $0.30/M input, $1.20/M output, and $0.06/M prompt-cache read for the standard ≤512K input-token row.
Q · 02 What happens above 512K input tokens? +
MiniMax lists a separate >512K standard row at $0.60/M input, $2.40/M output, and $0.12/M cache read. Its model page says the API supports up to a 1M-token context with a guaranteed minimum of 512K. MiniMax M3.1 Flash has no such step: one rate across its 1M window.
Q · 03 Does MiniMax M3 have Priority pricing? +
Yes. The Priority service-tier rows list $0.45/M input, $1.80/M output, and $0.09/M cache read for ≤512K, and $0.90/M input, $3.60/M output, and $0.18/M cache read above 512K.
Q · 04 What context window does MiniMax M3 support? +
MiniMax's text-generation docs list 1,000,000 tokens for MiniMax M3. The model page says the API supports up to a 1M-token context window with a guaranteed minimum of 512K tokens.
Q · 05 Is MiniMax M3 open weight, and is M3.1 Flash its replacement? +
MiniMax's self-hosting guide lists M3 as an open-weight model, alongside M2.7. M3.1 Flash, released on October 19, 2026, is a cheaper, faster sibling rather than a replacement: MiniMax says M3 "also remains available", and both stay on the price list.
Q · 06 Is MiniMax M3 multimodal? +
Yes. MiniMax describes M3 as native multimodal; the text-generation docs list multimodal chat input with text, image, and video content parts.
Q · 07 When was this price last checked? +
On 2026-10-23, against MiniMax's pay-as-you-go pricing page; all eight rates were unchanged since June. The page prints each row against struck-through reference rates twice as high and calls the reduction permanent, so the lower figures are the price.
Y
Reviewed by Yaroslav Vikhariev Founder · AI//COST · Prices checked by hand against platform.minimax.io · Last verified Oct 23, 2026
Methodology Report a correction More by Y.V.