GuShiio.com Article source: AI New List
Image source: Generated by AI
Hello friends, are you playing DeepSeek this Spring Festival?😎
Looking at this reasoning and writing ability, it is a world of difference compared with the model a year ago.
Domestic open source AI has really caused a global sensation this time. DeepSeek has become the first phenomenal AI product in the beginning of the year. Even my second uncle is chatting about “in-depth search” at the dinner table.🫡
This marks that the AI model is accelerating its penetration into ordinary people’s lives. In the past year, AI products have gradually shifted from “early adopters” to “just needed”. Many people have become accustomed to using AI to search for problems, learn new knowledge, assist in work, stimulate creativity, etc.
In order to help everyone understand and choose AI products more easily, starting from May 2024, the new list has launched the “AI New List-AI Product List”, including monthly and weekly lists. In previous lists, in addition to DeepSeek, which is rising rapidly in rankings, we can also find many popular AI products that are easy to use.
Starting in the New Year, if you don’t know which AI products are worth trying, you might as well take a look at our inventory. After all,”being able to use AI” is likely to be an important talent recruitment requirement in the future.
The following are the 2024 hot AI products and selected recommendations compiled by the “AI New List”. I hope they can help you seize the opportunity at work and make your life easier!
AI chat/general assistant
The 2024 “Hundred Models Battle” mainly focuses on multimodal and reasoning capabilities. Open source models are gradually catching up with closed-source models, and the competition for AI chat/universal assistants is becoming more intense. Starting from DeepSeek, we will see AI faster in the new year. evolution.
ChatGPT
OpenAI’s benchmark product ChatGPT supports long-text writing, code writing, multi-language translation, etc. The GPT-4o model upgraded in 2024 adds video understanding functions, advanced voice mode chat is more vivid, and the o3 model further strengthens logic capabilities. Recently, a new Deep Research feature has also been introduced. ChatGPT is suitable for complex tasks, but you need to subscribe to the Plus version (US$20/month) to unlock full features.
Official website: chatgpt.com/
bean bag
ByteDance’s multi-modal AI assistant Doubao is often at the forefront of the AI product list. It integrates text, image, music, video generation and other functions, has accurate Chinese understanding, and the upgraded AI voice can also sing. It is free and suitable for localization needs
Official website: www.doubao.com/chat/
DeepSeek
DeepSeek, a popular chicken in the AI circle at the beginning of the year, was developed by Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. The latest version of DeepSeek R1 is benchmark against OpenAI’s o1 model in performance and has outstanding reasoning capabilities, but the training cost is only one tenth of GPT-4o. Currently, DeepSeek’s web and App terminals are free to all users. It also supports API access to enterprise-level applications, and the response speed during peak periods is slightly slower.
Official website: chat.deepseek.com/
AI search
In 2024, AI search tools have been upgraded from “question and answer robots” to “intelligent information stewards”. They not only provide answers, but also integrate multi-source data, analyze trends, and even proactively predict demand.
Perplexity
Perplexity, the world’s leading conversational answer engine, can be described as an artifact for scientific research and industry analysis. It supports real-time networking + academic database retrieval. The answers are accompanied by source links, and supports questioning and result traceability. The free version is enough, and the Pro version (US$20/month) can call Deepseek R1, OpenAI o3-mini, Claude 3.5 Sonnet, Sonar and other models.
Official website: www.perplexity.ai/
Minita AI Search
Minita AI search, a must-have for professionals, is optimized for Chinese search, specializes in analyzing professional documents such as legal provisions and financial reports, supports direct analysis of PDF/web links, and is very accurate in identifying complex formulas. The answers are output in a structured manner, which can automatically generate summaries, key annotations, timeline sorting, tables, etc. Enterprise users can also upload internal documents to build a private knowledge base. Recently, I connected to the full-blood version of DeepSeek-R1. Free users can get 100 searches per day. After upgrading (39 yuan/month), it can be increased to 500 searches.
Official website: metaso.cn/
Tiangong AI Search
Tiangong AI is the first AI search product in China launched by Kunlun Wanwei. It supports multimodal searches such as images, documents, audio and video. The advanced model is suitable for academic research or in-depth financial analysis, and can generate brain maps with one click. Deepseek R1 deep thinking mode has been recently added. Search results are biased towards domestic sources, so attention should be paid to the timeliness of information. “Treasure Book” and “Color Pages” are its special functions. Creators can apply to become the manager of the Treasure Book and have the opportunity to earn advertising shares.
Official website: www.tiangong.cn/
AI companionship
AI emotional companionship products belong to the current hot Red Sea track. They mainly meet users ’emotional needs through role-playing and dialogue interactions. The interactive form has expanded from pure text dialogue to voice and video calls, and the product form has extended from software to physical hardware such as smart pets. The AI companion product market is expected to continue to grow rapidly in the next few years. According to Tencent Research Institute, the AI companion market will reach a scale of 100 billion in 2027.
Character.ai
Characteristic.ai is a representative product of AI accompanying the track. Its monthly active users reach 233 million worldwide. Its functions include custom character creation, multi-language support, personalized dialogue and Group chats mode. The paid version (US$9.99/month) can send more messages and have better character memory. In August 2024, Characteristic.ai was acquired by Google for US$2.5 billion. In addition, the platform triggered a public opinion crisis by providing extreme advice to young people, and subsequently launched parental control functions and strengthened filtering of inappropriate content.
Official website: character.ai/
Related reading: “Looking at the emotional fantasies and projections of millions of users from Character.AI Top 50 characters”
Hoshino
Hoshino is an AI content community owned by MiniMax. It is a domestically produced head AI emotional companionship product. The overseas version of Talkie also occupies a leading position in the United States, the United Kingdom and other markets. The product design integrates AI chat robots, card gameplay and UGC community. It supports voice continuous broadcasting and opening plot branches. Version 2.0 adds dynamic backgrounds, configured music, and multimodal gameplay. Pay (12 yuan/month) to get more exclusive benefits such as inspired recommendation responses.
Official website: www.xingyeai.com/
Loveydovey
LoveyDovey is an AI companionship application developed by South Korean company TainAI. It is mainly aimed at second-dimensional users and star-chasers. It mainly simulates real love experiences. Adult users can enhance their character’s goodwill. From acquaintances chatting to “marriage”, they have a strong sense of substitution and have been evaluated as “Japanese and Korean characters simulation ceiling”. Users need to consume virtual currency “jam” to chat with the AI character, charge per sentence (15 yuan can send about 100 messages), and also need to pay to view the AI character’s inner activities.
Official website: www.loveydovey.ai/en
Related reading: “Chongshang ranked first in South Korea on iOS, this South Korean AI chat software has become the new favorite of star-chasing and dreamers”
AI image
AI image products can automatically generate images based on text descriptions and reference maps, or modify and enhance existing images. At present, AI imaging products are widely used on both sides of B and C, and market demands for drawing design, retouching and photo have been verified.
Midjourney
Midjourney can be said to be the ceiling of AI artistic creation. Its artistic style covers a wide range of areas and is suitable for inspiration and creativity. The latest V6.1 version has better image/texture quality. In 2024, it will also launch an infinite canvas function that supports collaboration, Patchwork, personalized customization kinetic energy Moodboards, etc. Midjourney needs to be used through Discord or the official website. The US$30/month package generates unlimited pictures. Newbies need to learn basic prompt skills.
Official website: www.midjourney.com/imagine
Flux
Flux is the dark horse of the AI image open source model launched by Black Forest Laboratory in 2024. Its image clarity (8K resolution) and detail fidelity surpass Midjourney. The Flux Schnell version is the fastest to generate, with pictures in 5 seconds; the Dev version can be obtained and used on platforms such as Hugging Face; the Pro version is suitable for professional designers, and recently a fine-tuning API has been launched.
Official website: blackforestlabs.ai/
Jimeng AI
Jimeng AI is a one-stop AI creation platform under ByteDance. It is friendly to domestic users. The 2.1 model launched at the end of 2024 supports the generation of Chinese and English characters. It can generate Chinese posters with one click. It can also use a one-sentence command to modify picture styles, movements or expressions. It is suitable for quickly producing design materials. Free users get 88 points per day, which is enough for light use.
Official website: jimeng.jianying.com/ai-tool/home/
AI music
AI music products cover many fields such as tone cloning, Text To Speech, music generation, and sound editing. Among them, music generation products allow users to create complete music works with simple prompts, support multiple music styles and languages, and promote the application of AI music in advertising, games, podcasts and other scenarios.
Suno
Suno, the world’s leading AI music product, can generate complete songs with human voices through text. The V4 version can generate up to 4 minutes, with higher sound quality and more creative song structure. Free users are limited to 10 songs per day, and works for paying users (starting from US$10/month) are commercially available and are suitable for UGC content creation.
Official website: suno.com/home
ElevenLabs
ElevenLabs is a leader in the field of AI dubbing. It has mature tone cloning technology and supports multilingual emotional Text To Speech. It only takes 1 minute to clone sounds. Suitable for application scenarios such as audiobooks, virtual characters, and video dubbing. Free users get 10 minutes of text-to-speech time per month. The paid version starts at US$5/month and supports commercial use.
Official website: elevenlabs.io/
AI video
Sora, which had been waiting for almost a year, was finally officially opened to the public at the end of last year. However, by this time, domestic AI video generation products had caught up and greatly improved in terms of picture expressiveness, semantic understanding and controllability. AI video is moving from the model level to the application level. Whether it is UGC mass creation that is good at using AI to abstract, or film-level PGC and PUGC professional user creation, the popularity of AI video creation will continue to rise in 2025.
Keling AI
Keling AI is an AI creation platform launched by Quick Hand. It supports text/picture student film-level videos. It has outstanding motion control and physical simulation capabilities, and can extend videos to about 3 minutes. Recently, multi-picture reference and creative special effects functions have been added to the graphic video mode of the 1.6 model. Free users can receive 366 inspiration points per month (about 18 5-second videos are generated), which is slower to generate. The paid version (starting from 66 yuan/month) is faster, and you can also download watermarked videos.
Official website: klingai.kuaishou.com/
Conch AI video
Conch AI Video is an AI video generation product launched by MiniMax. It supports text/graphic videos and subject references, and is good at capturing and expressing the complex emotions of characters. The new model T2V-01-Director can use natural language or specific instructions to accurately control lens movement, and can also superimpose multiple lens movement methods at once. Free users receive 100 shells per day (3 videos can be generated), and paid members (starting from 68 yuan/month) receive 1000 shells per month.
Official website: hailuoai.com/video
AI 3D
3D content generation is the next AI competitive highland. It is currently in the early exploration stage. Some technology giants and startups have begun to launch 3D generation tools and platforms based on AI large models, which are expected to be used in game development, industrial design, film and television production, and e-commerce display. Implement applications in other fields.
TripoAI
Tripo, the AI 3D model generation platform launched by VAST, can generate 3D models in 10 seconds through text/single image/multiple image doodling, or even customized styles. The user interface is friendly and greatly reduces the threshold of 3D creation. Tripo can accurately restore the three-dimensional structure of objects, support material mapping optimization, and can also animate the model by binding bones. Free users earn 600 points per month (about 24 graphic 3D models), and more features start at US$19.9 per month.
Official website: www.tripo3d.ai/
Related reading: “A hat brings nearly 2.5 million yuan in revenue, and domestic AI 3D tools are beginning to help creators make money”
Rodin
Rodin, a 3D generation platform launched by ShadowEye Technology, supports Wensheng 3D and Tusheng 3D, and allows users to upload images from different angles. The system fuses these images to generate a more complete and accurate 3D model. Rodin is good at creating cartoon characters and can directly generate commonly used PBR maps. New users will receive 5 points, starting from US$15/month for the professional version.
Official website: hyper3d.ai/
AI programming
In 2024, AI programming tools have made significant progress in improving development efficiency, code generation and optimization. From assisted coding to independent development, even developers who cannot write code can try to create a simple application using AI.
AI programming tools are classified by level and field: paradite.github.io/ai-coding/? trk=public_post_comment-text
Cursor
Cursor is a representative new generation AI programming product that supports intelligent code completion, error prompts, real-time code generation, automated document generation and other functions. Compared with assistive tools such as GitHub Copilot, Cursor has significant improvements in automation and completion. It can seamlessly integrate large models into the development environment. The interaction is simple and friendly, and applications can be developed through chat. The free version has limited features and slow response times, while the Pro version ($20/month) runs offline.
Official website: www.cursor.com/
Bolt.new
Bolt.new is an AI-driven codeless development platform that can generate prototypes of web pages and Mini programs through dialogue. Relying on WebContainers technology, it not only generates corresponding code, but also integrates environment configuration, operation and deployment. All operations can be completed directly in the browser without additional development environment configuration. It is very friendly to novice developers and rapid prototyping. There are 1 million tokens for free users, and you can get more credits by paying (starting at US$20/month).
Official website: bolt.new/
AI Office
The application of AI in office scenarios includes document writing, PPT production, reading assistants, data analysis, meeting management, etc., which can help professionals reduce the time spent processing cumbersome tasks and improve work efficiency and decision-making accuracy.
NotebookLM
NotebookLM is a research and note taking tool launched by Google. Based on the Gemini 2.0 model, it can summarize and analyze uploaded content in various formats such as PDF documents, websites, YouTube videos, audio, and PPT, and supports interactive questions and answers. Among them, the “Audio Overview” function can automatically convert document content into conversational podcast audio, and users can also interact with AI anchors through the “Join Conversation” function. Free users can make up to 50 chat queries and 3 audio generation times a day, and Plus users have more quotas.
Official website: notebooklm.google/
Gamma
Gamma is an AI-driven presentation creation platform. Through simple text input and creative prompts, it can quickly generate PPT with professional design, including elements such as titles, text, pictures and charts. It supports real-time sharing and collaboration, and is suitable for project display and reporting. New users have 400 points, which can be generated 10 times for free. Plus (US$8/month) and Pro (US$15/month) users have unlimited number of creations.
Official website: gamma.app/
In addition to our selection of popular AI products, what other useful treasure AI products do you know?
