Dedicated Transit based on the Official API

In this era of openness and sharing, OpenAI has led a revolution in artificial intelligence. Now, we proudly announce to the world: We fully support all of OpenAI's models, as well as various domestic large models. Most excitingly, we are ready to unveil the more powerful and influential GPT-4-Turbo to the world!

  • Daily API Calls 100K+
  • Supported Large Model Number 100+
  • Number of Registered Members 120K+
home image

Our Services

GPT 3.5 Turbo

GPT-3.5 Turbo, based on the GPT-3 architecture, offers performance comparable to GPT-3 but at a lower cost, making it an ideal choice for developing various applications. It is suitable for a wide range of tasks, including text generation, semantic search, text summarization, translation, function calling, and more. Whether you are a developer, researcher, or business user, you can rely on the powerful capabilities of GPT-3.5 Turbo to easily build efficient and intelligent applications.

Other Large Models

Here, you will have the unique opportunity to experience domestic large models supported by China's top technology companies—Alibaba, Zhizhu, 360, iFlytek, Baidu, Tencent, and others. These models represent the latest technological breakthroughs in various fields within the country, covering artificial intelligence, natural language processing, image recognition, and more. They are not just the pinnacle of technology but also the pinnacle of intelligent services.

GPT-4 and 32K

GPT-4 is the latest version in the OpenAI GPT model series. Building on the features of GPT-3, it significantly improves in terms of model scale and comprehension abilities. GPT-4 has stronger generative capabilities and depth of understanding, capable of being applied in more complex scenarios and understanding more complex commands. Our site now also perfectly supports GPT-4-32K, providing users with a superior and smarter service experience.

DCDN Global Acceleration

Our interfaces use DCDN global acceleration to ensure that no matter where you are in the world, you can receive fast and stable services. DCDN (Dynamic Content Delivery Network) is a type of dynamic content distribution network that automatically selects the best route based on the user's geographical location, device type, network conditions, and other factors. It adjusts the content delivery strategy in real time to achieve optimal performance.

Official Billing Standards

In this smart world, you have the opportunity to shape your own future. We eagerly look forward to each of your innovations, as we join hands to build a new era of artificial intelligence together. We promise to strictly adhere to official billing standards without any hidden fees. Every request will be transparently charged according to official standards, ensuring you have a clear understanding of each expense, so there's no need to worry about any other issues. Rest assured and enjoy the convenience and joy brought by intelligent technology.

Data Security

In our Large Language Model (LLM) aggregation API platform, data security is at the core. We employ multiple security measures, including industry-standard data encryption, strict access control, and multi-factor authentication, to protect user data from unauthorized access. We also regularly conduct security audits and vulnerability scans to ensure the platform's security. Our goal is to provide a secure and reliable service platform that users can trust and use with confidence.

How to Start?

Quick Start with Dedicated Large Model API

Here are the summarized steps to get started. Please make sure you understand AI-related content beforehand.

  • 1 Register an account, and we will gift you $0.5 in trial credits.
  • 2 Copy your API KEY.
  • 3 Insert it into your program and replace the proxy domain with api.modelless.co
See detailed instructions
about image

Prices match official rates, unlimited RPM with no throttling (unlike similar sites' 3RPM limit)

Model Price Overview (Click to view the complete list and prices, refer to this link for accuracy)

Model Name Query Price Response Price Model Limit Notes
suno-v3 $0.3 /usage - 1K tokens The most powerful AI music generation model in history
suno-v2 $0.2 /usage - 1K tokens The most powerful AI music generation model in history
gpt-3.5-turbo $0.0015 /1k tokens $0.002 /1k tokens 4K/tokens Suitable for quickly answering simple questions.
gpt-3.5-turbo-16k $0.003 /1k tokens $0.004 /1k tokens 16K/tokens Suitable for quickly answering simple questions with more words.
gpt-3.5-turbo-1106 $0.001 /1k tokens $0.002 /1k tokens 16K tokens Latest model, most up-to-date data, lower price.
gpt-3.5-turbo-instruct $0.0015 /1k tokens $0.0020 /1k tokens 8K tokens Fine-tuned model, suitable for special scenarios.
gpt-4 $0.03 /1k tokens $0.06 /1k tokens 8K tokens The most powerful model, more capable.
gpt-4-1106-preview $0.01 /1k tokens $0.03 /1k tokens 4K tokens Latest model, maximum context memory of 128K, maximum output of 4K.
gpt-4-all $0.03 /1k tokens $0.06 /1k tokens 32K tokens Multifunctional GPT-4 model, integrated with various capabilities.
gpt-4-gizmo $0.03 /1k tokens $0.06 /1k tokens 32K tokens Official GPTs model, focused on specific application scenarios, compatible with all GPTs plugins.
gpt-4-vision-preview $0.01 /1k tokens $0.03 /1k tokens 8K tokens Latest model, multimodal.
gpt-4-32K $0.06 /1k tokens $0.12 /1k tokens 32K tokens The most powerful model, more capable, with more words.
gemini-pro $0.0015 /1k tokens $0.003 /1k tokens 8K tokens Google's latest model, offering more complex language understanding and generation capabilities.
gemini-pro-vision $0.003 /1k tokens $0.006 /1k tokens 8K tokens Google's latest model, providing more complex language understanding and generation capabilities.
bing $0.01 /1k tokens $0.01 /1k tokens 8K tokens Microsoft Bing's official model.
dall-e-3 ---- HD1024x1024: $0.08 per image ---- GPT's latest drawing model. (Supports high definition)
midjourney ---- fast: $0.23 per image ---- Midjourney drawing model, available via API. Click to see how to use
mj-chat ---- $0.3 per image ---- Midjourney drawing model, chat form.
stable-diffusion ---- $0.3 per image ---- SD drawing model, chat form.
claude-1.3-100k $0.008 /1k tokens $0.008 /1k tokens 100K tokens Claude's latest model.
claude-2 $0.008 /1k tokens $0.008 /1k tokens 100K tokens Claude 2 official model.
claude-3-opus-20240229 $0.06 /1k tokens $0.06 /1k tokens 100K tokens Claude 3, the latest official model.
tts-1 ---- $0.015 /1K characters ---- GPT text-to-speech model.
tts-1-hd ---- $0.03 / 1K characters ---- GPT text-to-speech model, high definition.
SparkDesk ¥0.018 /1k tokens ¥0.018 /1k tokens 8K tokens iFlytek's SparkDesk v3.1.
ERNIE-Bot ¥0.012 /1k tokens ¥0.012 /1k tokens 8K tokens Baidu's ERNIE Bot model.
ERNIE-Bot-turbo ¥0.008 /1k tokens ¥0.008 /1k tokens 8K tokens Baidu's ERNIE Bot model.
ERNIE-Bot-4 ¥0.12 /1k tokens ¥0.12 /1k tokens 8K tokens Baidu's ERNIE Bot v4.0 model.
chatglm-pro ¥0.01 /1k tokens ¥0.01 /1k tokens 8K tokens Zhipu's ChatGLM Pro model.
chatglm_std ¥0.005 /1k tokens ¥0.005 /1k tokens 8K tokens Zhipu's ChatGLM Standard model.
chatglm-trte ¥0.004 /1k tokens ¥0.004 /1k tokens 8K/32K tokens Zhipu's ChatGLM model.
chatglm-turbo ¥0.005 /1k tokens ¥0.005 /1k tokens 8K tokens Zhipu's ChatGLM Turbo model.
hunyuan ¥0.1 /1k tokens ¥0.1 /1k tokens 8K tokens Tencent's HunYuan large model.
qwen-plus ¥0.02 /1k tokens ¥0.02 /1k tokens 8K tokens Alibaba's Qwen Plus large model.
qwen-max ¥0.02 /1k tokens ¥0.02 /1k tokens 8K tokens Alibaba's Qwen Max large model.
qwen-turbo ¥0.008 /1k tokens ¥0.008 /1k tokens 8K tokens Alibaba's Qwen Turbo large model.
text-embedding-ada-002 $0.0001 /1K tokens $0.0001 /1K tokens 8K tokens GPT-emb vector model, 50k+/rpm.

FAQ

The site calculates fees based on the number of characters used, billing for the resources consumed during your requests to the GPT resources. For streaming requests where the interface does not return the consumption amount, the server uses a tokenizer to calculate the consumption accurately for each model's corresponding encoder (e.g., gpt-3.5/gpt-4 series models use cl100k_base encoder, while text-davinci-003 model uses p50k_base encoder). The model's input and output are calculated separately according to the standard, and the amount of characters consumed is deducted accordingly. For synchronous requests, the consumption amount returned in the interface is used directly.
First, this is not a bug. Since it's hard to precisely predict the length of language model outputs, your account balance might occasionally become negative. If your balance reaches 0 or goes negative, we will send a notification email to your account, and you might not be able to use most services on the site, including certain web services and APIs. Please recharge your account promptly to restore access.
We provide a dedicated line for users with stable usage levels (subject to a certain recharge threshold). After contacting customer service to become a Direct level member, you will have exclusive access to a dedicated line that is not shared with other users, ensuring 100% stability and efficiency.
There are certain thresholds to become an agent, including a requirement for a specific volume of users. Users can contact the administrators to apply to become an agent, and we will help agents deploy a sub-site. Agents can then earn a corresponding proportion of the revenue generated from the recharges promoted by their site.
We support invoicing; please contact customer service. The recharge amount is added to the account balance, and if invoicing is required, an additional 11% tax will be applied.