MiniMax H3MiniMaxText-to-Video
MiniMax's omni-modal audio-video generation system unifies text, image, video, and audio context to create videos up to 15 seconds at 2K with native stereo sound for advertising, e-commerce, and film production.
MiniMax
MiniMax APIs for M2.7, H3 & Hailuo
Build coding agents and multimodal video workflows with the MiniMax models currently available through LinkModel.
Provider overview
About MiniMax
MiniMax develops model families across language and multimodal generation. M2 models focus on coding, professional work, and agentic workflows; H3 unifies text, image, video, and audio context for audio-video generation; and Hailuo models provide established text-to-video and image-to-video workflows. LinkModel only presents models that appear in its live catalog.
Explore MiniMax language and video APIs
Choose M2.7 for coding and long-running agents, H3 for general-purpose multimodal video generation, or a listed Hailuo model for efficient visual production. This catalog only shows MiniMax models currently available through LinkModel.
4 models listed
MiniMax H3 and Hailuo video APIs
Compare MiniMax H3 and Hailuo models for text-to-video, image-to-video, multimodal references, motion control, resolution, duration, and model-dependent native audio.
MiniMax H3
MiniMax Hailuo2.3MiniMaxText-to-Video
MiniMax's upgraded video model built on Hailuo 02, enhancing complex body motion, facial micro-expressions, and physical realism, with expanded style support including anime, ink wash, and game CG — better quality at the same price.
MiniMax Hailuo02MiniMaxText-to-Video
MiniMax's cinematic AI video generation model built on its proprietary NCR architecture with 2.5x efficiency gains, supporting native 1080p output, extreme physics simulation, and precise instruction following, ranked among the top global video models.
MiniMax language and agent APIs
Use the live MiniMax M-series models for software engineering, tool-driven agents, professional work, and complex multi-step delivery.
MiniMax M2.7MiniMaxChat
MiniMax's self-evolving 230B MoE model (10B active) with native Agent Teams, dynamic tool search, and end-to-end engineering delivery — MLE Bench Lite 66.6% trails only Opus 4.6/GPT-5.4, built for SRE, ML research, and office automation.
Compare MiniMax models and API pricing
Start with MiniMax M2.7 for coding and long-running agents, MiniMax H3 for general-purpose multimodal audio-video production, or MiniMax Hailuo 2.3 for established motion-focused and stylized video workflows. Recommendations only appear when the matching model is live in LinkModel's catalog.
MiniMax H3 and Hailuo video APIs
Best for
A general-purpose multimodal generation model for complex references, native stereo audio, flexible duration, and high-resolution commercial video workflows.
Key capabilities
Multimodal references · native stereo audio · up to 15 seconds · up to 2K
Official source ↗PricingView pricing
Best for
An established video family for controlled human motion, facial expression, stylized output, and efficient text-to-video or image-to-video production.
Key capabilities
Human motion · facial expression · stylization · text or image input
Official source ↗PricingView pricing
MiniMax language and agent APIs
Best for
MiniMax's current language model for real-world software engineering, complex tool use, agent teams, and multi-step professional delivery.
PricingView pricing
What can you build with MiniMax APIs on LinkModel?
Move between agentic language work and production video generation while managing credentials, usage, and billing from one account.
Coding and operations agents
Use M2.7 for software engineering, log analysis, debugging, tool-driven automation, and long-running tasks that must preserve a coherent execution plan.
Commercial multimodal video
Use H3 for advertising, branding, e-commerce, product design, and other workflows that combine text, image, video, or audio references with native audio-video output.
Stylized and motion-focused content
Use a listed Hailuo model for character movement, facial performance, anime, illustration, game-CG, and efficient product-motion iterations.
Agent-to-video production pipelines
Use M2.7 to plan, structure, or automate a campaign workflow, then send the approved creative direction to a live MiniMax video model.
Why run MiniMax models through LinkModel?
Use supported MiniMax language and video models alongside other providers without adding another account or billing workflow.
One key across language and video
Use one LinkModel API key for supported MiniMax agent and video models in the live catalog.
MiniMax alongside other providers
Compare or combine MiniMax with other language and video families without maintaining a separate platform account for each provider.
Centralized usage and billing
Review language-token and video-generation usage from one LinkModel account while selecting the right model for each workload.
Integrate MiniMax APIs
Use one LinkModel credential to select a live MiniMax model, submit the matching request, and monitor usage from the same account.
Authenticate with one API key
Create a LinkModel API key and use it to authenticate requests for the MiniMax models enabled for your account.
Choose a live model and request schema
Open the M2.7, H3, or Hailuo model page to confirm its current identifier, supported task, required inputs, and pricing units.
Submit and track the workload
Handle a language response directly or poll an asynchronous video task until the generated asset is ready.
MiniMax model guides
MiniMaxVideo GenerationAI ModelsGuides
MiniMax H3 API Pricing: Pay-as-you-go Costs
2026-08-27
MiniMaxPrompt EngineeringVideo GenerationGuides
MiniMax H3 Video Prompt Guide: Examples
2026-08-27
MiniMaxAPIVideo GenerationGuides
MiniMax H3 API Guide: Video Generation Tutorial
2026-08-27
hailuo vs seedanceminimax hailuoseedance 2.0ai video comparisonvideo generation api
Hailuo vs Seedance: Which AI Video Model Should You Use?
2026-07-10
FAQ
Frequently asked questions.
Sources and editorial information
Editorial review
Model capabilities and selection guidance are reviewed against the official MiniMax sources listed here.
Editorial content reviewed: September 1, 2026
MiniMax and its model names belong to their respective owners. LinkModel is an independent API aggregation platform and does not imply a formal relationship.