Tencent launches Hy4 Flash multimodal model with one-million-token context

What happened
Tencent introduced Hy4 Flash, a new multimodal AI model, on November 16, 2026. The model features a one-million-token context window and is designed to be faster and more affordable than previous versions. Hy4 Flash is available through Tencent Cloud and AI Gateway, with a 50% promotional discount running until November 29, 2026. The model supports advanced functions such as reasoning, tool calling, and image input.
Key details
- Hy4 Flash has 224 billion parameters, including 64 billion Engram parameters.
- Maximum input and output limits are 960,000 and 64,000 tokens, respectively.
- The model is accessible on Tencent Cloud, AI Gateway, and several other Tencent platforms.
- The 50% discount on API usage is available through November 29, 2026.
Why it matters
Businesses and developers using Tencent Cloud and AI Gateway can now access a large-context multimodal AI model at significantly reduced costs, lowering expenses for high-volume productivity and agent tasks. The promotional pricing and expanded capabilities make it more feasible for organizations to process large documents, images, and complex tasks at scale. This release gives users a new option for cost-effective AI deployment, while those needing higher intelligence for complex software projects may still require the full Hy4 model.

