MiniMax

MiniMax API & Models on GPTProto

  • New Models on Day One
  • Free Auto-Routing
  • No Surcharges

Build coding agents, analyze long documents, generate short-form video from text or images, synthesize multilingual speech, and create consent-based voice clones with MiniMax models on GPTProto. Compare live Model IDs, route-specific pricing, and task fit before opening the right model page.

Models

ModelInput → Output
Minimax M31.05M$0.48 / $0.96 per 1M$0.10 / $0.10 per 1M
Input: TextInput: ImageInput: Document
Output: Text
Minimax M2.5205K$0.24 / $0.96 per 1M$0.30 / $0.02 per 1M
Input: TextInput: Document
Output: Text
Speech 2.6 Hd—— / $60.00 per 1M—
Input: Text
Output: Audio
Hailuo 2.3 Fast—$0.17 per time—
Input: Image
Output: Video
Hailuo 2.3 Pro—$0.44 per time—
Input: TextInput: Image
Output: Video
Hailuo 2.3 Standard—$0.25 per time—
Input: TextInput: Image
Output: Video
Hailuo 02 Standard—$0.25 per time—
Input: TextInput: Image
Output: Video
Hailuo 02 Pro—$0.44 per time—
Input: TextInput: Image
Output: Video
Hailuo 02 Fast—$0.09 per time—
Input: Image
Output: Video
Speech 2.5 Turbo Preview—— / $36.00 per 1M—
Input: Text
Output: Audio
Speech 2.5 Turbo Preview Voice Clone—$0.50 per time—
Input: TextInput: Audio
Output: Audio
Speech 02 Turbo—$0.002 per time—
Input: Text
Output: Audio
Speech 02 Hd—$0.0082 per time—
Input: Text
Output: Audio
Speech 2.5 Hd Preview Voice Clone—$0.50 per time—
Input: TextInput: Audio
Output: Audio
Speech 2.5 Hd Preview—— / $60.00 per 1M—
Input: Text
Output: Audio

Pricing

Cost calculator

Multi-turn agent with cached context.
$0.48 / 1M$0.00072
$0.96 / 1M$0.000768
$0.096 / 1M$0.000288
$0.096 / 1M$0.0024
Cost per request$0.004176

Top up

GPTProto vs official pricing.
20% off
Requests
You pay$100
You receive$100.00

Save$8.91 (8%)vs MiniMax official

About MiniMax

MiniMax is a global AI foundation model company developing systems for language, video, speech, music, and image workflows. On GPTProto, the supported MiniMax catalog centers on M-series language routes for coding and agentic work, Hailuo routes for short-form video generation, and Speech routes for text-to-speech and voice cloning. Model names can describe a family, generation, quality tier, or operation, so use the live route labels and documentation rather than assuming similarly named endpoints behave the same way. This MiniMax models collection updates as GPTProto adds, changes, or removes supported routes; it does not represent every model released by MiniMax.

Why Use MiniMax with GPTProto?

Use one account to compare and call supported MiniMax routes across language and creative-media workflows.

One Key Across Modalities

Use one GPTProto key for supported MiniMax text, video, and speech routes, following each route’s documented request format.

One Shared Balance

Fund MiniMax and other supported providers from one prepaid balance instead of managing a separate balance for each platform.

Compare Real Billing Units

Compare token, per-generation, and audio billing units on live cards before choosing a model for production.

Switch Without Another Account

Move between compatible routes or providers under the same account, while keeping route-specific endpoints and parameters explicit.

How to Access MiniMax Models

Get started with a MiniMax API route through GPTProto.

  1. 1

    Create a GPTProto Account

    Sign up and open the dashboard to manage your prepaid balance, API keys, and usage.

  2. 2

    Generate an API Key

    Create a key in the dashboard and store it securely in your application or server environment.

  3. 3

    Call a Live MiniMax Route

    Choose a live Model ID from the catalog, then follow that model page’s endpoint, parameters, and request format.

MiniMax API FAQ

Which MiniMax models are available on GPTProto?

The live catalog above is the source of truth for supported MiniMax routes, Model IDs, modalities, status, and current pricing. It may include only part of MiniMax’s official product line and updates as GPTProto availability changes.

Which MiniMax model should I use for coding or agents?

Use MiniMax-M3 when long context, tool use, multimodal input, or sustained agent work is central to the task. MiniMax-M2.5 remains an alternative for established text-only coding and office workflows where its displayed input price is more important. Check each live card for the current route status.

Which MiniMax model should I use for video generation?

Start with the input and output badges. Hailuo 2.3 Pro is the Picker choice for quality-first image-to-video work, while Hailuo 02 Pro covers the text-to-video path shown on GPTProto. Use the live catalog to compare Fast, Standard, and Pro route pricing before generating.

Which route should I use for text-to-speech or voice cloning?

Speech 2.6 HD is the Picker choice for high-fidelity multilingual TTS, while Speech 02 Turbo is a lower-cost alternative for latency-sensitive speech. Voice cloning uses a dedicated route rather than the ordinary TTS call; only clone a voice when you have the speaker’s authorization and the necessary rights.

Do I need a separate MiniMax API key?

No. Use your GPTProto API key for the MiniMax routes listed on this page. Different modalities may use different endpoints, fields, or asynchronous workflows, so follow the selected model’s GPTProto documentation instead of assuming one request format works everywhere.

How are MiniMax prices, discounts, and rate limits determined?

Billing units vary by route: language models may bill by tokens, video routes by generation or output settings, and speech routes by their documented audio unit. Discounts should compare matching models and billing units. RPM, TPM, and concurrency limits also vary, so check the live model card and route documentation before production use.

How does GPTProto process and store API data?

GPTProto processes API inputs and outputs to complete requests and may retain them temporarily for transaction processing and error debugging. Technical usage metadata may also be logged. Content sent to an upstream provider is subject to that provider’s policies. Review the current Privacy Policy before sending sensitive data.

Can MiniMax outputs be used commercially?

As between you and GPTProto, you retain rights in your inputs and outputs, subject to MiniMax and other applicable third-party terms. Commercial use, watermarking, and content restrictions can vary by model and route. You remain responsible for your prompts, reference media, voice permissions, and compliance with applicable policies and law.

Start with MiniMax

One GPTProto API key for MiniMax and other providers.

Get API Key