MiniMax Launches H3 Multimodal AI Model; Shares Surge on Cost Claims

MiniMax Debuts H3, Promising Cheaper 2K Video Generation

Shares of Chinese AI startup MiniMax (HK:0100) jumped on Friday after the company introduced H3, a new open-source multimodal artificial intelligence model. H3 can process and unify text, images, video and audio in a single context, and is capable of generating 2K-resolution video clips up to 15 seconds long with native stereo sound.

The company explicitly targets commercial applications, including advertising, e-commerce, gaming, product design and content creation. MiniMax claims that H3 delivers industry-leading cost efficiency, pricing 2K video generation at less than one-third of comparable mainstream models per second, while 768p output costs half as much as rivals charge for 720p video.

MiniMax also plans to release the model's weights in the coming days, subject to applicable laws and regulations, in order to support developers and improve compatibility with various AI hardware platforms. The launch intensifies an already fierce race among Chinese AI firms to publish increasingly powerful open-source models to compete with US-based challengers.

How MiniMax's H3 Pricing and Open-Source Strategy Shake Up China's AI Video Market

The open-source gambit: a developer land grab?

MiniMax's decision to go open-source reflects a strategy used by several Chinese AI companies to quickly build an ecosystem and disrupt incumbents. By offering model weights, it aims to attract developers who might otherwise rely on more expensive, closed-source US models. However, the caveat that the release is subject to applicable regulations introduces real uncertainty; Chinese authorities have previously tightened controls on the export of advanced AI technology.

Pricing war in AI video generation

If MiniMax's headline cost claims hold up, they could accelerate a race to the bottom in generative video pricing. For high-volume, cost-sensitive creative sectors such as advertising and e-commerce, the promise of 2K video at under a third of current market rates could reshape content production economics. Rival Chinese and US models will be forced to respond, potentially squeezing margins across the sector.

Technology leap or marketing flash?

H3 introduces features like Omni Contextual Representation and In-Context Regeneration, which MiniMax says improve instruction following, multimodal understanding and high-resolution generation. Without independent benchmarks or third-party validation, however, it remains unclear how these claims translate into real-world performance. The sharp stock move signals that investors are betting on the narrative at least as much as the technology itself.

What the H3 Launch Means for Content Producers and AI Developers

For content production teams: Evaluate H3's pricing structure against current video generation costs. If the model matches claimed output quality, advertising agencies and e-commerce platforms could cut video production budgets significantly by integrating an API-based workflow.

For AI developers: Monitor the open-source weight release closely. Early access could provide a competitive edge in fine-tuning video models for niche applications, but factor in the risk that regulatory delays may postpone availability.

For investors: The surge reflects optimism around MiniMax's cost leadership, but note that the company's claims remain unverified. Watch for rival responses from both Chinese and US firms—a quick matching of price points could erode MiniMax's advantage.