AI Category AI Model Developer Model Version Key Features Use Cases Service Price Offered by AIFUNS
Text AI ChatGPT OpenAI - GPT-4o mini: the most basic model (stronger than domestic AI).
- GPT-4o: the fast, stable version of the standard model.
- GPT-4o with canvas: the fast, stable version of the standard model with an added local-editing (canvas) feature.
- GPT-4: the standard model.
- GPT-o1-mini: the fast, stable version of the advanced model.
- GPT-o1-preview: the advanced model, with the strongest reasoning and intelligence; it can make mistakes on simple questions and has a limited number of uses (40 times per week).		 - GPT-4o mini: handles simple tasks (such as asking about the weather); can search the web and upload files.
- GPT-4o: handles slightly more complex questions (such as looking up book references); can search the web and upload files.
- GPT-4o with canvas: local editing is suitable for long-form writing; can upload files.
- GPT-4: more complex questions (middle/high school math calculations); can upload files.
- GPT-o1-mini: complex questions, a complement to GPT-o1-preview; that is, after using GPT-o1-preview to think through and answer, use GPT-o1-mini for local edits to save time and usage count.
- GPT-o1-preview: complex questions (research, questions with large amounts of file data).		 Chatbots, content creation, code writing, language translation, educational tutoring, conversational assistants, text summarization, Q&A systems, content generation. 125 RMB/month
1. Includes all services
2. No VPN needed for access within China
3. Usage count is double the official usage count
Claude Anthropic - Claude 3 Haiku: the basic model.
- Claude 3 Opus: the advanced model; limited by its version, its reasoning and intelligence are not as strong as 3.5 Sonnet.
- Claude 3.5 Sonnet: the mid-tier model, the latest version.		 - Claude 3 Haiku: handles simple questions; can upload files.
- Claude 3 Opus: the advanced model; limited by its version, its reasoning and intelligence are not as strong as 3.5 Sonnet; can upload files.
- Claude 3.5 Sonnet: the mid-tier model, the latest version; can upload files.		 Chatbots, content creation, code writing, language translation, educational tutoring, conversational assistants, text summarization, Q&A systems, content generation. 125 RMB/month
1. Includes all services
2. No VPN needed for access within China
3. Usage count is double the official usage count
Audio AI Suno Suno AI - V3.5: the latest model, can generate music up to 4 minutes long. High-fidelity text-to-audio generation, supporting multiple languages and emotions. Speech synthesis, audio content creation, audiobook production. 85 RMB/month
1. Includes all services
2. Access within China requires a VPN
3. 500 generations/month
Fish Audio Hanabi AI Inc Can perform speech synthesis based on an uploaded human-voice audio file Simulates a human voice, voice output of text up to 500 characters Speech synthesis, audio content creation, audiobook production. Domestic; purchase on your own
Image AI Midjourney Midjourney - V6.1: the latest model, can generate images in various styles. Available on the official site Text-to-image, image reference, transformations, local editing, etc.; relatively stable Advertising, printing, film and TV, art 240 RMB/month
1. Includes all services
2. Access within China requires a VPN
3. 15 hours of fast generation time
Stable Diffussion StabilityAI - V3: the latest model, can generate images in various styles. Locally deployed Users need to find a suitable trained model according to their own needs to generate the desired images; requires skill and luck Advertising, printing, film and TV, art Assistance: remote deployment fee is 300 RMB/time, customer must provide the equipment
Tutorial: 100 RMB
DALL·E OpenAI - V3: the latest model, can generate images in various styles (cannot generate ultra-realistic portraits). Available on the official site Text-to-image, image reference, transformations, local editing, etc.; relatively stable Illustrations for presentations Included in the ChatGPT fee
Video AI Pika Pika Labs Pika 1.5: can generate 5-second videos Text-to-video, image-to-video, special videos (such as squashing, shattering, etc.) Film and TV 280 RMB/month
1. Includes all services
2. Access within China requires a VPN
3. 2000 credits/month
Luma Luma AI V1.6: can generate 5-second videos Text-to-video, image-to-video Film and TV 240 RMB/month
1. Includes all services
2. Access within China requires a VPN
3. 150 times/month
Runway Runway ML - Gen-3 Alpha Turbo: can generate 10-second videos Text-to-video, image-to-video, special features (such as keyframes, etc.) Film and TV 280 RMB/month
1. Includes all services
2. Access within China requires a VPN
3. 2250 credits/month
Hailuo Conch AI - abab-video-1: can generate 6-second videos Text-to-video, image-to-video, special features (such as character facial expressions, etc.) Film and TV Domestic; purchase on your own
Kling No information available - Kling 1.5: can generate 10-second videos Text-to-video, image-to-video, special features (such as keyframes, etc.) Film and TV Domestic; purchase on your own