Tencent uses artificial intelligence to convert text to video Hunyuan Video revealed that it has been published as open source; The interesting thing is that this AI is released right when we are during the 12-day campaign of OpenAI; In this campaign, it is possible that the company will release the AI video maker Sora to the public after months of waiting.
According to the Hunyuan Video website, Tencent’s new artificial intelligence is a new and advanced model that creates high-quality videos through text prompts. According to the company, Hunyuan Video’s 13 billion parameter model is the most powerful open source video production model available.
Artificial intelligence to convert text to video from Tencent
“We present Hunyuan Video, a new open-source model whose video production performance is comparable to, if not better than, leading closed-source models,” Tencent said in its official announcement. In the table below, you can see the benchmarks of Tencent’s new model compared to other models:
This Chinese company claims that its new model is better than advanced models such as Runway Gen-3 or Luma 1.6 based on the results of expert evaluation. Of course, with this new model, you can only produce 5-second videos. Also, each video takes about 15 minutes to produce.

Tencent’s new model uses a multimodal linguistic model (MLLM)-based text encoder and outperforms previous state-of-the-art models with a text alignment rate of up to 68.5%. The text alignment rate measures how faithful the AI model is to your prompt and how well the output matches what you wrote. Tencent also says the AI has a 96.4 percent visual quality score.

Hunyuan Video is open source on GitHub; But if you want to use this model on the main site, you need to buy credits to be able to produce videos; You also need 15 credits for each video. Each 150 credit pack costs $9.99. Of course, you can also try this artificial intelligence for free at FAL.ai.
RCO NEWS