A clearer view of AI.
THE INDEPENDENT AI.

FRONTIER LAB / LLM · WORTH A LOOK

MiniMax

Multi-modal model company: MiniMax M3 LLM, H3 video, Speech 2.8, Music 3.0, plus Agent and Design apps.

Visit minimax.io ↗@MiniMax_AI on X ↗

What MiniMax does

Multi-modal model company: MiniMax M3 LLM, H3 video, Speech 2.8, Music 3.0, plus Agent and Design apps. Builds frontier coding/agentic LLMs (M3, 1M context) and media models, which it sells through an API and products: MiniMax Code/Agent (@MiniMaxAgent), MiniMax Design/Hailuo (@Hailuo_AI) and Talkie. Says it has over 200M users.

Good for

  • Low-cost LLM API for agents and coding
  • AI video (H3) for ads
  • Voice and music generation for content

Try this in 10 minutes

  1. Low-cost LLM API for agents and coding. Pick one real example from your own business and try just that.
  2. AI video (H3) for ads. Put the result next to how you do it today. Is it faster, cheaper or better?
  3. Voice and music generation for content. Note what you had to fix by hand before you'd show it to a customer.

Mini-tasks based on common small-business uses for this tool.

Honest take

Some claims not independently checked

  • Its models also show up on fal/Together/OpenRouter; often cheaper than US models.
  • Also covers @MiniMaxAgent and @Hailuo_AI (MiniMax Design).

Pricing

API pay-per-token (e.g. MiniMax-M3 $0.30 input / $1.20 output per M tokens at current 50% discount); app plans Unknown.

Prices as listed by the maker when we last checked (2026-10-23); usually in US dollars. Check the website for current prices.

Latest update

  • 2026-10-23: MiniMax M3.1 Flash open weights released (428B total, 23B active)
  • : MiniMax M3.1 Flash open weights released on Hugging Face and ModelScope (428B total, 23B active per token), aimed at agentic tasks, coding and cybersecurity (73.7% DeepSWE v1.1, 61.7 OSWorld V2, 82.2 CyberGym); day-0 support from Together AI, vLLM/Inferact, SGLang and NVIDIA NeMo fine-tuning.
  • : MiniMax announced its next model, MiniMax-M3.1-Flash (nicknamed 'Space Bunny'), is coming; no release date, specs or pricing in the post yet.
  • : Released MiniMax M3.1 Flash: 1M context, $0.15/$0.60 per M tokens; MiniMax claims ~3x faster and 70% cheaper agent tasks; Artificial Analysis Intelligence Index 42 but high hallucination rate.

Is MiniMax worth it? Quick answers

Is MiniMax worth it?

MiniMax is worth a look if it matches a job you already do (Frontier lab / LLM). Its models also show up on fal/Together/OpenRouter; often cheaper than US models. It is a paid tool, so test it on one real task before committing. Some of the maker's claims have not been independently checked.

How much does MiniMax cost?

API pay-per-token (e.g. MiniMax-M3 $0.30 input / $1.20 output per M tokens at current 50% discount); app plans Unknown.

What is MiniMax used for?

Multi-modal model company: MiniMax M3 LLM, H3 video, Speech 2.8, Music 3.0, plus Agent and Design apps. Typical uses: low-cost LLM API for agents and coding; AI video (H3) for ads; voice and music generation for content.