ChatGPT на русском и другие нейросетив одном сервисе

ChatGPT is a bot that uses the GPT-3.5 model to communicate with users.

GPT-3.5 - This is a chatbot based on the original GPT-3.5 language model from OpenAI introduced in 2022. The model was trained using a corpus of texts from the Internet and human learning system. Often its data is limited to 2021, which can affect the accuracy and correctness of the answers. We recommend checking whether the model's answers are accurate and using it for popular questions.

GPT-4 - It gives higher quality answers, but works slower and consumes more words when used. It answers thoughtfully and in detail (if asked) and is less prone to "hallucinations". At the moment, this is the most advanced text neural network from OpenAI

GPT-4-Turbo. The key difference from GPT-4 is that it works 2-3 times faster and the cost of use is about 2.5 times less. At the same time, the quality of the answers remains close to the level of GPT-4.

In this case, you can renew your subscription without waiting for the next month. You can take the same tariff plan or another one. The limit will be expanded immediately, and the next charge will be made a month after the last (just made) payment.

The number of words is calculated by the formula: number of tokens / 15. A token is a single word or punctuation mark.

• Entered Text;
• Text generated by the neural network;
• Text of previous messages necessary for the neural network to take into account the context of the dialogue for future answers (context).
• Attached files containing text (e.g., documents).

By default, the neural network in a specific chat remembers and processes the last 2000-16000 characters (250 - 2000 words), depending on the settings and the model used.

When history is turned off (the "settings" button to the left of the input field), the neural network will not spend energy on accounting for upper messages - only the entered and generated text will be counted.

In GPT-4 mode, about 15 times more words are spent than actually processed.

When generating images, the number of words from the text sent to the bot is spent, plus a fixed number of words for generating the image itself:

• For Midjourney 5.2 (per one request = 4 images) — 3200 words;
• For Stable Diffusion (per one request = 4 images) — 3200 words;
• For DALL·E 3 (per one request = one image) — 3600 words;
• For DALL·E 3 HD (per one request = one image) — 7200 words.

We use the latest paid versions of neural networks. Including Chat GPT from OpenAI. This guarantees the stability and speed of the neural network.

Yes, you can ask a question or write a technical problem in the Telegram chat in the support section and we will answer.

We accept payment by Visa, MasterCard, MIR bank cards, as well as Apple Pay and Google Pay.

You can cancel your subscription at any time. To do this, go to the "subscription" section in the settings and click "cancel subscription".

To disable auto-renewal, you need to go to your personal account and click the "Cancel auto-renewal" button. When paying by SBP, auto-renewal is not set up, so you do not need to disable auto-renewal.

After disabling auto-renewal, the already paid subscription will continue to work until the end of the term or until the word limit is reached.