The author ported Tencent's video world model Hy-Nomas to native ComfyUI nodes, with the DiT used as a regular ComfyUI MODEL sampled by the native KSampler, and it runs on just 12GB VRAM.
Why it matters: The author ported Tencent's Hy-Nomas to native ComfyUI nodes and provides measured speed and prompt-writing guidance under 12GB VRAM, which can be used to assess the feasibility of running a world model locally.
Tencent Hunyuan has open-sourced its world model Hy-Nomas, which supports free interaction using text prompts and camera trajectories and has long-horizon generation capability with world memory. The model provides lightweight LoRA adaptation for vertical domains such as robotics; after 8-step distillation it can generate 720p@24FPS video on a single RTX 4090. The weights are now available on Hugging Face and GitHub, along with an online Demo.
Why it matters: Tencent Hunyuan open-sources world model Hy-Nomas, supporting text and camera trajectory control; 720p video generation runs on a single RTX 4090.
Tencent News validated the news short drama format with "Burmese Den of Demons" and "AI Recreates 100 New Types of Scams"; the former has over 2.2 million views across the web, while the latter launched 8 episodes in one month, exceeded 2.6 million views across the web, and appeared on subway and public security screens. The article outlines the collaboration model co-created by procuratorial and judicial bodies such as the Supreme People's Procuratorate and the Supreme People's Court together with media, and summarizes three creative paths: institutional co-creation, AI industrial processes, and communication design. The author also points out unresolved issues such as a low rate of premium works, the absence of professional journalism norms, and the higher risk of AI hallucinations in news scenarios.
Tencent Hunyuan open-sourced the interactive world model Hy-Nomas on October 21, 2026, releasing weights, inference code, and training and data annotation code, along with an online demo and the Hy-Nomas-Bench benchmark.
Why it matters: Tencent Hunyuan open-sources an interactive world model in which text and camera trajectories can drive long video generation segment by segment, with a benchmark and single-GPU configuration.
In mid-October 2026, the short-drama businesses under ByteDance and Tencent密集 introduced new policies. Volcano Engine announced incentives of up to 200,000 yuan for premium AI animated short dramas; Hongguo Short Drama launched an AI creation plan with a 3 million yuan guarantee fund, with the single top prize exclusively receiving a 1 million yuan guarantee; and its 2.0 rule canceling the series-scale coefficient also officially took effect in October.
The WeChat Open Platform and the WeChat Channels team jointly launched the "Official Account Novel Adaptation Micro Short Drama: Phase One" incentive campaign, selecting 10 works in the first phase; the platform has confirmed adaptation intentions with the authors. If producers and authors reach cooperation during the campaign period, complete the adaptation, and prioritize distribution within the WeChat ecosystem, both parties can each receive an additional 5% cash incentive on revenue during the priority distribution period, with no upper limit.
Broadcast Network launched pre-reorganization after creditor Xi'an Guanghua Communication applied on the grounds of 5.1133 million yuan in matured debt unpaid; the Xi'an Intermediate People's Court has accepted the case and designated a liquidation group as interim administrator, while publicly recruiting reorganization investors, with registration closing at 18:00 on November 8, 2026 and a 10 million yuan deposit required. China Broadcasting Network Shanxi Company held a "Hundred-Day Sprint, Win the Year" progress meeting on October 10, calling for accelerating the deep integration of "one network for the whole province" and improving the three-tier business data reporting mechanism.
In the first half of 2026, China Literature's short drama and AI comic drama businesses brought in over 430 million yuan in revenue, more than three times the same period last year; as of the China Literature Creation Conference on October 16, more than 660 AI comic dramas had surpassed 10 million views, 107 of which exceeded 100 million.
Tencent launched the "Green Bubble Short Drama" mini-program on WeChat, taking on Hongguo Short Drama with a social fission logic. Short drama creators may face a split in content production. Green Bubble relies on WeChat's social graph for social recommendations, with free viewing and ad monetization, and content provided by third-party producers. Hongguo had 168 million daily active users as of July 2026, relying on ByteDance's feed-based ad delivery system. The two platforms may form a stratified market, and creators will need to split content production according to platform preferences—one script cannot easily serve both platforms.
Why it matters: Green Bubble enters short drama via WeChat's social chain, contrasting with Hongguo's user-acquisition logic; creators may need to split content production according to platform preferences.
MMLab@HKUST of HKUST, Tencent Video, and the University of Hong Kong propose Adaptive Reward Routing, which, on the DiffusionNFT framework, recomputes multi-reward routing for audio-video diffusion models in real time during training rather than fixing it once from the base model.
The WeChat Short Drama Content Center launched the "Lüpaopao Short Drama" mini program in late September, featuring the slogan "Free good dramas, watch with one tap," with three sections: Free Theater, Swipe, and Me, adopting a joint distribution and advertising model. The company says that after integration, user time spent is 5 times that of the overall WeChat short drama market, and high-quality new dramas with per-user watch time of ≥10 minutes can receive up to 100% revenue incentives. The article compares this with Hongguo Short Drama's scale as of July 2026 of 168 million daily active users and 356 million monthly active users, and notes that WeChat enters with advantages in close-tie social networking and no download requirement, but still has shortcomings in exclusive content and experience.
Why it matters: WeChat Steps In to Launch a Free Short Drama Mini-Program; Comparing Market Share, Watch Time, and Monetization Data with Hongguo Reveals a New Round of Changes in the Short Drama Platform Landscape.
The Doubao phone assistant published a list of 151 apps that refuse GUI screen automation, including WeChat, Taobao, Meituan, DeepSeek, and Kuaishou; users can still install and use them manually. Tencent quietly launched the Green Bubble short drama mini program, released by the WeChat short drama content center, with a "Friends are watching" recommendation slot on the homepage, targeting ByteDance's Hongguo short dramas. A Saudi national engineer at OpenAI was arrested and fired for hiding an encrypted recording device in a meeting room at the San Francisco office, and faces up to five years in prison.
ByteDance's overseas short drama app PineDrama surpassed 100 million downloads and 7 million daily active users within 9 months of launch, making it the only new product among the 7 standalone short drama apps with cumulative global downloads exceeding 100 million that has been live for less than a year. It replicates Hongguo's free-plus-ads model, with Indonesia and Brazil as its main battlegrounds; the two countries together account for about 82% of cumulative downloads, while the US accounts for only 6.7%. Tencent, meanwhile, fills the gap with its WeChat mini-program Green Bubble short dramas and announced a co-distribution policy offering up to 100% revenue incentives for average viewing time of more than 10 minutes per person.
Starting October 16, WeChat has launched a new round of micro short drama mini program incentives. Co-distributed high-quality new dramas can receive up to 100% revenue incentives, consisting of 90% cash share and 10% advertising credits, with advertising credits only usable for ad placements. Previously, during the 7-day National Day holiday, WeChat's Green Bubble Short Drama accumulated 43,000 sets of ad materials, with daily average material volume increasing by 57% compared to the week before the holiday.
Duanju Zixishi reviewed the AI interactive short drama sector, noting that the focus of competition has shifted from whether options exist to how much audience input can rewrite the story, and used two dimensions—interaction method and generation timing—to distinguish pre-made branching products from real-time generation products. Castloop, under Yuanxiang Technology, has launched more than 1,000 works, covering more than 20 countries, with a cumulative total of more than 50,000 registered users and peak daily active users exceeding 10,000. Its self-developed interactive narrative memory model is built on the XVERSE-Ent foundation, with main storylines and endings pre-set and constrained while branch details are generated in real time.
Zhipu GLM-5.3-Highspeed has appeared in gray-release feedback on the Zcode community, with some 3.15.1 subscribers randomly receiving this high-speed generation version; the general release scope and API pricing have not yet been confirmed. Microsoft has released Microsoft-Decision-1, which is post-trained based on Qwen3.5-9B and completes binary judgment, candidate selection, and scale scoring in a single forward pass. It is available for public preview on Foundry and hosted access on OpenRouter.
Tencent ARC released Multics on Hugging Face, achieving scalable multi-person video generation through layered communication. The repository includes the complete 4-player Multics DiT checkpoint, configuration, Wan2.2 VAE, UMT5 text encoder, and tokenizer; the same checkpoint supports 6-, 8-, and 10-person generation without retraining. The weights are distributed under the Apache 2.0 license.
Why it matters: Tencent ARC open-sources the multi-person video generation model Multics, where the same weights can scale to 10 people, letting creators evaluate generation options for multi-person scenes.
Tencent's Lü Paopao short dramas increased user acquisition spending during the 2026 National Day holiday period, benchmarking against ByteDance's Hongguo short dramas, as the free short drama market sees head-on competition. Hongguo short dramas have 168 million daily active users, exceeding the combined total of four long-video apps; Lü Paopao relies on the WeChat ecosystem, using mini-programs and social distribution for differentiated defense.
Why it matters: Reviewing the decade-long video game between Tencent and ByteDance, comparing the two paths of free short dramas, the competitive landscape and differentiated opportunities for Lü Paopao and Hongguo are clearly presented.
The AI short drama industry shifted from capacity shortage to oversupply within half a year, with content prices dropping from 5,000 yuan per minute to just a few hundred yuan. In September, Douyin Group's native AI short dramas averaged 1,010 new releases per day. Industry companies are migrating in three directions: upward into AI mid-to-long dramas, such as "Northern School Notes" which peaked at 11.3% market share and exceeded 1 million yuan in first-week revenue sharing; into new AI interactive content, where DRAMAGO achieved a 40% monthly retention rate within 3 months of entering the market; and outward into overseas markets, where Kunlun Wanwei's AI short drama platform exceeded $65 million in monthly revenue.
Tencent's review of four short drama trials shows they all got stuck at the channel level. WeChat's Green Bubble is WeChat's first time directly producing short dramas, produced by WeChat Short Drama Content Center, not as a standalone app but growing within the WeChat ecosystem. Green Bubble's Q4 incentive policy uses average playback duration per person ≥10 minutes as the threshold. High-quality new dramas can enjoy up to 100% revenue incentive (90% cash share + 10% advertising bonus), basic incentive 85%, and only 50% if duration is not met.
Why it matters: Tencent's review of four short drama trials shows they all got stuck at the channel level. Green Bubble is held by WeChat officially as the product entity and uses average playback duration incentives to filter content, shifting its positioning from a channel to an aggregator.
Wenyu Jiazhi Guan reviewed Tencent's four previous short-drama attempts, noting that Shifen Theater, WeChat mini programs, Huoxing Short Drama Club, and others all stalled at the channel level and failed to build brand recognition or aggregate traffic. WeChat Green Bubble, launched in September this year, is produced by the WeChat short-drama content center, follows the free IAA route, and its Q4 incentive policy is tiered by average playback time per user, with up to 100% revenue incentives. The article compares Hongguo's 168 million daily active users with the WeChat ecosystem's 427 million monthly active users of short-drama mini programs, arguing that Green Bubble's value lies in bringing scattered traffic into an official entrance rather than directly challenging Hongguo.
Wenyu Jiazhi Guan reviewed Tencent's four previous short-drama attempts, concluding that Shifen Theater, WeChat mini programs, Huoxing Short Drama Club, and the WeChat short-drama mini program all stalled at the channel level and failed to build brand recognition or aggregate traffic. WeChat Green Bubble, launched in September this year, is produced by the WeChat short-drama content center, follows the free IAA route, and its Q4 incentive policy is tiered by average playback time per user, with up to 100% revenue incentives. The article compares Hongguo's 168 million daily active users with the WeChat ecosystem's 427 million monthly active users of short-drama mini programs, pointing out that scattered traffic makes it hard to gain industry influence.
The author used UniversalModder to stuff DOOM into Skyrim, creating SkyDoom, and demonstrated how to mash up any two games you own; UniversalModder is a skill for Claude Code, Codex, or Gemini CLI—just point it at two games and state your requirements.
Tencent's Hunyuan team, in collaboration with Fudan University, released the open-source video model Prism, which natively generates 2K video with audio—approximately 10 seconds at 24fps with 48 kHz audio. It only supports image-to-video, requires an 80 GB GPU for 720p, and has no quantized version yet. The code is under the Tencent-Hunyuan GitHub organization, preview weights are on a personal Hugging Face account, and Tencent has not issued an official announcement.
Wireflow ranks first in the 2026 Hunyuan3D v3 tool comparison, placing concept image generation and 3D conversion on the same node canvas and supporting REST endpoints and MCP tool calls.
Tencent Hunyuan releases Hunyuan Image 3.5 Preview (Hy Image3.5 preview), positioned as a high-value professional-grade image generation model. It supports text-to-image and image-to-image generation, up to 5 reference images per upload, up to 2K resolution output, and multi-turn conversational creation.
Why it matters: Tencent Hunyuan releases a preview of its professional-grade image generation model, providing 2K output, API pricing of 0.15 yuan per image, and internal blind test comparison results, which can be used to assess cost and image generation capability.
Tencent ARC released WorldCrafter-Base on Hugging Face, including base transformer weights along with a camera adapter and LoRA, for use with WorldCrafter inference code.
Tencent ARC released self-contained WorldCrafter-Fast weights on Hugging Face, supporting image-to-video, text-to-video, and an interactive image-to-video demo, with inference possible without WorldCrafter-Base files.
Why it matters: WorldCrafter-Fast has released standalone weights, with sampling step configurations for both image-to-video and text-to-video made public, making it easier for creators to evaluate integration costs.
Tencent Hunyuan, in collaboration with Fudan University and Zhejiang University, proposed Prism, a dynamic sparse attention framework for natively training 2K joint video-audio generation models. It organizes token sequences into spatiotemporal macro-blocks and dynamically allocates block shapes based on local content, achieving a 2.5x training speed improvement over full attention with higher generation quality.
Tencent ARC open-sourced GAE-D64-1B, a temporal DiT with about 1B parameters (hidden_size=[768, 2048], depth=[28, 6]) that operates on a 64-channel geometry-native latent space, supports 672×378, V=81, and uses metric Plücker camera conditioning with a Qwen3-0.6B text encoder.
Chinese open-source models have nearly unanimously shifted to MoE architecture, with Kimi K2, MiniMax M2, and Qwen3 all adopting this approach, aiming for sustainable deployment at controlled costs. Open-source activity expanded to multimodality and agents starting February 2025, with StepFun's Step-Audio-R1.1 voice model outperforming closed-source models, and Tencent open-sourcing Hunyuan Video and Hunyuan 3D.