/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

← → days · ↑ ↓ browse · Enter similar · o open

Tencent launches Hy Image 3.5 Preview, saying that during internal testing the AI image model achieved results on par with ByteDance's Seedream 5.0 Pro

Tencent Holdings Ltd. launched its latest and best image-generation model on the same day as rival Alibaba Group Holding Ltd. kicked off its AI conference …

Bloomberg

Context & Ripple Effects

Tencent has been extending its generative-media stack, from text- and image-to-3D tools in 2025 to the open Hy4 Preview foundation model in August 2026. Hy Image 3.5 narrows that broader push to image generation, where Tencent is publicly using ByteDance’s Seedream 5.0 Pro as its quality reference point.

Alibaba has been in the category since its 2023 Tongyi Wanxiang enterprise beta. Releasing an image-model preview alongside Alibaba’s AI conference makes image quality a direct point of comparison among major Chinese AI platforms.

First-order effects

  • Tencent adds a new preview model to its image-generation lineup and says its internal testing places it on par with ByteDance’s Seedream 5.0 Pro.
  • ByteDance’s Seedream 5.0 Pro becomes the explicit benchmark Tencent uses to frame Hy Image 3.5’s performance claim.

Second-order effects

  • Alibaba faces a sharper comparison at its AI conference: its generative-image positioning is measured against Tencent’s stated Seedream benchmark as well as its own earlier Tongyi Wanxiang work.
  • Enterprise buyers evaluating Chinese generative-media tools gain another vendor claim to test on image-specific workflows, shifting competition from general model scale toward output quality in a defined format.

Third-order effects

  • If leading vendors continue releasing format-specific models and publishing rival-relative results, image generation is likely to become a distinct competitive layer rather than a feature subsumed by general-purpose foundation models.
  • The durable advantage will depend less on a single internal benchmark than on whether image models can be incorporated into each platform’s wider creative-tool and enterprise offerings.

The trend: Chinese AI platforms are increasingly competing through format-specific generative models, with image-quality benchmarks becoming a visible battleground alongside broad foundation-model releases.

Discussion

  • @tencenthunyuan Tencent Hy on x
    Hy Image3.5 preview is live. 🚀 Professional-grade image generation, +30% win rate in human eval vs Hy Image3.0 Both Text to image & Image to image available. Up to 2K. Better Consistency. API: https://console.tencentcloud.com/ ... Priced for everyone. $0.024 per image on Tencent …
  • @tencenthunyuan Tencent Hy on x
    Share some genius cases generated by Hy Image3.5 preview. 😎
  • @tencentglobal @tencentglobal on x
    Tencent Hy Image 3.5 preview is now available, bringing stronger image generation and consistency across both text-to-image and image-to-image creation. See what's new and try it out below. 👇
  • @onsoloai @onsoloai on x
    OnSolo Limited Time Membership Perk Members get free Hy-Image-3.5-Preview generations for the duration of the event. Basic: 3,000 images Premium: 10,000 images Super: 28,000 images Ultra: Unlimited Event period: September 18 to October 6. All quotas expire at the end of the event…
  • @sameerdhara Sameer on x
    Just saw Tencent drop Hy Image 3.5 preview. Looks like a real jump from 3.0. They're claiming a 30% win rate in human evals, better consistency, text-to-image and image-to-image, up to 2K. $0.024 an image and they don't charge for the reference photo. Two weeks free on OnSolo and…
  • @alextalksai Alex on x
    Most image models still can't spell. I gave Hy Image3.5 preview a billboard with 6 text layers and an 85mm product shot in the same session. It generated me the best poster result for the upcoming best fragrance for winter. Nailed both. First try. Tencent's new model went live to…
  • @heyrohitai Rohit on x
    First test: commercial/product visuals. I wanted to see how well it could handle realistic materials, lighting, reflections, textures, and overall photography quality. The goal: make an AI-generated image feel much closer to a real commercial shoot.
  • @tencentai_news @tencentai_news on x
    A professional image model that doesn't cost like one. $0.024 per 2K image. Meet Hy Image3.5 preview. > up to 5 reference images in, and the meter only runs on what comes out >multi-turn editing with earlier edits in context > text, layout and 2K consistency, the parts that break…
  • @aiwithshoaib @aiwithshoaib on x
    Tencent built an image model that actually spells its own text right. Hy Image 3.5 preview is live in @Miora_Design today. I got early access before the crowd. Three things I threw at it: → a poster packed with text → a product image edit → a full UI screen Here's what came back …
  • @alex_inspira Alexander Inspira IA on x
    @OnSoloAI The quality of its images surprises for its price. If we want to generate content without raising costs, it is definitely worth a try.