Stay free if you only need api access to all endpoints and model playground. Upgrade if you need dedicated capacity and sla. Most solo builders can start free.
Why it matters: English-only reasoning benchmarks still favor Anthropic and OpenAI frontier models
Available from: Pay-as-you-go
Why it matters: Cinematic-quality video generation lags Runway and Kling in side-by-side tests
Available from: Pay-as-you-go
Why it matters: Self-hosting M1 requires substantial GPU memory — not a laptop workload
Available from: Pay-as-you-go
Why it matters: Documentation is Chinese-first with translated English pages
Available from: Pay-as-you-go
MiniMax is used for coding, agentic workflows, multimodal generation, video creation, speech, music, and developer API access. Its website highlights MiniMax M3, Hailuo AI video generation, Speech, Music, and open platform access.
MiniMax M3 is the company's coding and agentic model referenced on the website. It uses MiniMax Sparse Attention, supports up to a 1M-token context window, and is presented for coding, agents, and long-context workflows.
The official pricing pages visible in the provided content list Token Plan subscriptions at $20/month, $50/month, and $120/month, with monthly MiniMax M3 token usage of about 1.633B, 5.053B, and 9.796B tokens respectively. MiniMax also references free or freemium access, credits, video packages, audio subscriptions, and API usage pricing.
Yes. The website lists Hailuo 2.3 as MiniMax's current video generation model and includes it among the flagship models alongside MiniMax M3, Speech 2.8, and Music 2.6.
Based on the visible content, MiniMax is broader than many category-specific competitors because it combines coding, agentic workflows, long-context processing, video, speech, music, image generation, and API access.
Start with the free plan — upgrade when you need more.
Get Started Free →Still not sure? Read our full verdict →
Last verified March 2026