Tencent Hunyuan says its Hy Image3.5 preview is now live with text-to-image and image-to-image generation, output of up to 2K, and a stated Tencent Cloud API price of $0.024 per generated image. In a post on X, @TencentHunyuan also describes the model as “professional-grade” and reports a “+30% win rate” in human evaluation versus Hy Image3.0.

The announcement does not provide the evaluation methodology or examples, so the comparison is a result reported by Tencent rather than an independently verified performance test. The “+30% win rate” wording also does not establish whether the figure is a relative increase, a percentage-point difference, or another metric.

What Hy Image3.5 preview offers

The preview supports two listed workflows:

  • Text to image: generating an image from a text prompt.

  • Image to image: generating an image using a reference image as input.

Tencent Hunyuan also says the model can produce images at up to 2K resolution and offers “Better Consistency.” The post does not define that term or provide technical specifications, supported file formats, or benchmark details.

Where to try Hy Image3.5

The announcement points developers to the Hy Image3.5 preview API on Tencent Cloud. It says the API charges $0.024 per generated image and that reference images are not charged. These are the pricing details stated in the announcement; the linked page should be checked for current terms and any additional requirements.

The post also lists two services offering two weeks free, specifically through Miora and OnSolo. It does not provide further eligibility, account, or other terms for those offers.

For now, the available information describes a preview announcement rather than a full technical release. Readers can use the API or the named services to explore the model, but the post alone does not establish how Hy Image3.5 performs across different prompts, image formats, or production workloads.

Sources