MiniMax

MiniMax closed a $300 million Series B extension in July 2025 at a $4 billion valuation, led by Shanghai state-owned capital through Shanghai STVC Group.

The company previously raised funding from Alibaba Group and Tencent-affiliated vehicles in Series A and B rounds. Other investors include HongShan Capital Group, Hillhouse Investment, IDG Capital, Yunqi Capital, Vitalbridge Capital, and GL Ventures.

MiniMax has raised approximately $1.15 billion in total funding since its founding.

MiniMax builds a suite of multimodal AI models for text, images, video, audio, and music generation. The company's language models include the MiniMax-Text-01 with 456 billion parameters and a 4 million token context window, the hybrid-architecture MiniMax-M1 that requires 70% fewer computational resources than competing models, and the M2.5 and M2.5 Lightning models optimized for enterprise AI agents with performance approaching state-of-the-art while maintaining cost efficiency.

For consumers, the Hailuo Video app allows users to enter text prompts like "a hummingbird diving through a rainforest" to generate 6-10 second video clips in 30 seconds. The Talkie app enables users to create custom AI characters with specific appearances, personalities, and voices for text or voice conversations.

Developers access these capabilities through APIs on platform.minimax.io. The Speech-02 model converts text to speech in 40 languages and can process up to 200,000 characters for audiobooks or game dialogue. The Music-1.5 model composes 4-minute songs with vocals and instruments, including Chinese folk instruments, while Hailuo-02 generates 1080p video clips with physics consistency and camera controls.

The platform also offers real-time multimodal APIs with sub-second latency for conversational agents and Video Agent templates for storyboard creation with a single click.

MiniMax runs a dual-sided model with consumer mobile apps and enterprise API services. On the consumer side, a freemium approach keeps basic usage ad-supported, with premium subscriptions removing ads and adding unlimited messaging for $9.99 monthly.

The enterprise model uses usage-based pricing by modality. The M2.5 model is priced at $0.15 per million input tokens and $1.20 per million output tokens, with M2.5 Lightning at $0.30/$2.40 per million tokens. Text processing costs $0.40 per million input tokens, video generation runs $0.10 per 6-second clip, and audio synthesis scales up to 20 million characters monthly through subscription tiers.

For enterprise AI agent deployments, MiniMax estimates M2.5 costs approximately $0.15 per task versus $3.00 for Claude, with four always-on agents costing roughly $10,000 per year. Developers can also bring their own API keys from other providers to reduce costs.

MiniMax trains proprietary models across all major modalities rather than licensing from third parties. This vertical integration tightens quality control and can raise margins, though it requires substantial upfront R&D investment in compute infrastructure and model training.

As more users generate data that can be used for training, model performance can improve, and enterprise customers often expand usage across multiple modalities once integrated.

Alibaba's Qwen models, Tencent's Hunyuan, ByteDance's Doubao, and Baidu's Ernie are a primary source of competition due to distribution reach and cloud infrastructure. ByteDance cut API pricing to $0.02 per million tokens and uses TikTok's user base for consumer AI apps.

These platforms bundle AI capabilities across existing ecosystems (Alibaba through AliCloud and Taobao, Tencent via WeChat and gaming, ByteDance through TikTok and enterprise tools). Their vertical integration enables cross-subsidization and customer acquisition at scale that pure-play AI companies find difficult to replicate.

MiniMax competes directly with other Chinese AI unicorns including Moonshot, Zhipu, Baichuan, and DeepSeek in the race to build frontier models. DeepSeek's recent R1 model release has intensified price competition across the sector, pushing peers to cut token pricing or open-source model weights.

In the consumer AI companion space, MiniMax's Talkie ranks among the top three globally alongside Character.AI and Replika in a $120 million market growing 88% year-over-year. These platforms compete primarily on character quality, conversation naturalness, and content moderation policies.

OpenAI, Anthropic, and other US-based AI companies represent indirect competition, particularly for enterprise customers and developers building global applications. However, regulatory restrictions and data sovereignty concerns limit their direct presence in the Chinese market, reducing head-to-head competition for domestic providers such as MiniMax.

MiniMax's Music-1.5 model targets the $30 billion global music production market by enabling 4-minute song generation with vocals and instruments. Use cases include professional music production, advertising soundtracks, and game audio.

The Hailuo-02 video model's native 1080p output and cinematic controls target the $150 billion digital video advertising market. Real-time multimodal APIs with sub-second latency are applicable to gaming, livestreaming, and enterprise collaboration where responsiveness matters more than perfect quality.

Speech-2.5's support for over 40 languages addresses global voice-over, dubbing, and call center markets worth approximately $50 billion combined.

Partnerships with creator platforms like Envato's VideoGen, Captions, and fal.ai put MiniMax models in front of millions of paid creators who often switch between tools. These integrations create a path from indirect usage to direct API billing and may lower customer acquisition costs.

The M1 language model is available through Together AI's GPU cloud, letting startups access the model without building their own infrastructure. The Model Context Protocol server and Video Agent workflows provide SaaS vendors prebuilt methods to embed multimodal generation.

MiniMax runs a program in South Korea that partners with film festivals and universities, used as a template for entering creator-heavy markets where US incumbents have limited brand recognition. The company already serves users across 200+ countries, which can support localized expansion.

Cross-border platform integrations offer an indirect path into North America and Europe despite potential geopolitical restrictions on direct market entry.

Regulatory constraints: Geopolitical tensions between China and Western markets could limit MiniMax's ability to expand internationally or to access global cloud infrastructure, constraining growth to domestic and aligned markets and giving competitors scale advantages in global markets.

Price compression: DeepSeek's price cuts have driven industry-wide price reductions across Chinese AI companies, with token costs falling to $0.02 per million. This margin pressure could make it difficult for MiniMax to achieve profitability while investing in the compute infrastructure needed to train competitive models.

Platform dependence: MiniMax's consumer apps face potential restrictions on major app stores and social platforms, while enterprise customers may prefer to work directly with cloud providers such as Alibaba or Tencent, which offer integrated AI services alongside core infrastructure.