Tencent 3D Rendering AI Model InstantMesh Unveiled, WiMi Accelerates Vertical Applications


A few days ago, Meta (META) released a new large language model, Llama 3. The model contains both 8 billion (8B) and 70 billion (70B) parameters.

It is reported that along with the release of Llama 3, Qualcomm (QCOM) said that it has entered into a cooperation with Meta, and in the future, the optimized Llama 3 can be directly run on smartphones, PCs, VR/AR headsets and automobiles equipped with Snapdragon platforms and other terminals.

Around the AI field, Qualcomm has built the AI Hub, which provides about 100 optimized AI models, which can facilitate developers to quickly develop as well as land the corresponding AI capabilities. Qualcomm said the partnership aims to empower OEMs and developers, further develop end-side AI, and drive the popularization of generative AI capabilities.

3D rendering AI models InstantMesh online

Tencent (TCEHY) followed suit by releasing InstantMesh, a new AI model that renders 3D objects using still photos.

InstantMesh is said to be an upgrade of Tencent’s old Instant3D framework, which combines a multi-view diffusion model with sparse view reconstruction of 3D objects based on the Large Reconstruction Model (LRM) architecture. Currently, Tencent has open-sourced this AI model InstantMesh model on Hugging Face so that developers can test its generation efficiency.

In addition, Tencent has launched a results preview application where modelling enthusiasts can add photos and preview their 3D renderings. According to research data, InstantMesh outperforms SV3D on Google Scanning Objects (GSO) and OmniObject3D (Omni3D) orbital views.

Currently, the curtain of the new decade-level technological and industrial revolution of AI has been opened, with phenomenal breakout products such as ChatGPT and SORA appearing, and software giants such as Microsoft (MSFT) and Adobe (ADBE) launching commercialized products.

WiMi’s comprehensive support for AI iteration

Driven by big models, WiMi Hologram Cloud (NASDAQ: WIMI) has been actively following the global AI model development boom. At present, WiMi actively takes the development of AI as the primary strategic choice, integrating AI big model, 5G, cloud computing, artificial intelligence, image rendering engine, holographic projection and other diversified industries and technologies, to promote the landing of AI industrial applications and ecological construction, which will inject a source of ‘AI power’ into the digital economy.

It is worth mentioning that the development of large model technology represented by ChatGPT provides a smarter “brain” for humanoid robots. WiMi has increased its investment in R&D to ensure the core competitiveness of its technology and products, and based on the complete AI+ vision technology system of “full-stack technology R&D capability + layout of technology routes in all fields”, WiMi has continued to carry out the R&D and technical research on robot vision and multi-modal interaction large models, etc., and continued to empower the high-quality development of the humanoid robot industry. We will continue to carry out R&D and technical research on robot vision and multimodal interaction models to continuously empower the high-quality development of the humanoid robot industry.

Artificial Intelligence is an important focus point for the development of new productivity, WiMi will accelerate the application of the AI ecosystem to a whole new level, deeply analyze the key role of AI big models in promoting industrial upgrading and transformation, explore innovative cooperation modes, open up the key barriers of AI technology empowering the industry’s innovation and development, and accelerate the promotion of the incubation and landing of big models in key areas. It will bring unprecedented intelligent experience to the public and open a new chapter of AI intelligent life.


The year 2024, as the first year of AI native application, coupled with years of technical accumulation and product operation capabilities of global manufacturers, has achieved a subversive and innovative application of large models and opened up the chain from evolution to realization of large models. If we say that now is the “wind mouth” of the big model, certainly there is a long journey ahead, and at this point, WiMi has seized the opportunity to run in the forefront.

In the future, large models will continue to lead a new round of artificial intelligence change. The industry generally believes that the big model is the core engine towards AGI general artificial intelligence, which will bring another leap in artificial intelligence, or become the focus of global technology competition. It is hoped that major technology companies will continue to attack the core technology, plough into the AI industry, bring more people a convenient, comfortable and technology-oriented living experience, help develop an open, win-win and sustainable AI industrial ecosystem, and truly bring AI into thousands of industries and households.