How much does the AI chat cost?

Your first request is free — try any of the 8 models without paying. After that, you pay only for the tokens you actually use, straight from your account balance: shorter requests cost less. You'll see the cost of each request as you type, and your balance can never go negative.

Do I need to sign up to use the AI chat?

Your first request is free, no sign-up needed. To keep chatting with any model, create a free MyKreaTool account and add funds to your balance; each request is billed by the tokens it uses.

Does the chat support file uploads?

Yes. You can attach PDF, TXT, DOC, CSV and JSON files, plus images (PNG, JPG, WebP). The AI reads the content and tailors its answer to your file.

Tools MyKreaTool

AI Chat

AI Chat with Top Models

Pick an AI model below to start chatting. Each model comes in several versions for different tasks.
Pay per token, only for what you use.

Choose an AI model

DeepSeek

Coding & analysis
pay per token

≈ ₽ / 1M

Gemini Flash

Google · fast AI
pay per token

≈ ₽ / 1M

GPT OSS 120B

OpenAI · open-weight 120B
via Groq · fast

≈ ₽ / 1M

ChatGPT

OpenAI · top model
for complex tasks

≈ ₽ / 1M

Claude

Anthropic · precision
and long-form writing

≈ ₽ / 1M

Gemini Pro

Google · long
context · powerful

≈ ₽ / 1M

Grok

xAI · fresh data
from Elon Musk

≈ ₽ / 1M

Premium

Council

All AIs at once →
Claude distills the best answer

≈ 14 191 ₽ / 1M

—

AI can make mistakes. Do not use it for medical or legal decisions.

AI model catalog: 7 models in multiple versions

MyKreaTool brings the world's best AI models together in one place: ChatGPT, Claude, Gemini, Grok, DeepSeek and GPT OSS. Each model comes in several versions, from fast and budget-friendly to the most powerful. Below, we break down each version's strengths and when to use it, so you can pick the right model for the job. Switch models without hopping between websites, and compare their answers side by side.

Free and budget-friendly AI models

DeepSeek (V4 Flash, V4 Pro)

A next-gen model that shines at coding, math and logic. Large context window, low price.

V4 Flash — fast and cheap, with up to 1M tokens of context. Best for: large documents, drafts and routine work where speed and cost matter.
V4 Pro (preview) — DeepSeek's strongest reasoning and coding, with long outputs. Best for: complex reasoning and large amounts of code where quality beats speed.

Gemini Flash (3.5, 2.5, 2.0)

Fast Google models for writing, translation, summaries and Q&A.

3.5 Flash — the newest and smartest of the fast models, strong at coding and agentic tasks. Best for: everyday work with near-top accuracy at Flash speed.
2.5 Flash — previous generation, cost-efficient. Best for: high request volumes on a budget.
2.0 Flash — an early budget model. Best for: simple bulk tasks at the lowest cost.

GPT OSS 120B

OpenAI's open-weight model (120B) on the Groq API — very fast responses and strong multilingual support.

120B — OpenAI's open-weight model on Groq, with the best speed-to-price ratio among open models. Best for: conversation, multilingual tasks and content creation when you want an open model that's fast.

Paid AI models

ChatGPT (5.5, 5.4, 5.2) Premium

OpenAI's flagship models for complex multi-step tasks, coding and professional content.

5.5 — the newest: fewer mistakes on sensitive topics (law, medicine, finance), strong math and multimodal skills. Best for: smart everyday chat and complex tasks.
5.4 — the previous generation, with Thinking and Pro modes. Best for: complex tasks when you prefer the familiar previous model.
5.2 — an earlier GPT-5 generation. Best for: basic tasks and saving money on the flagship line.

Claude (Opus 4.8, Sonnet 4.6, Haiku 4.5) Premium

Anthropic models known for precision on long texts, coding and analysis.

Opus 4.8 — the most powerful: top-tier reasoning and coding, with adaptive thinking depth. Best for: the hardest reasoning, large codebases and deep analysis.
Sonnet 4.6 — fast, with a great balance of quality and price. Best for: your everyday go-to for coding, writing and analysis at a mid-range price.
Haiku 4.5 — the fastest and most affordable. Best for: high volume and low latency on simple tasks.

Gemini Pro (3.1, 3, 2.5) Premium

Powerful Google models with the longest context, built for analyzing large amounts of data and multimodal tasks.

3.1 Pro (preview) — the current flagship: frontier-level reasoning, advanced math and coding. Best for: the hardest tasks when you need frontier quality.
3 Pro — the previous-generation Pro. Best for: high-quality reasoning and coding on the prior release.
2.5 Pro — the stable, fully supported GA release with long context. Best for: production work that needs stability.

Grok (4.3, 4.20) Premium

xAI's model with long context (up to ~2M tokens) and strong agentic capabilities.

4.3 (beta) — the newest: native video understanding and an improved architecture. Best for: multimodal tasks and long-context analysis.
4.20 Reasoning — thinks step by step before answering. Best for: math, logic, debugging and nuanced analysis.
4.20 — a fast flagship with agentic tool calling. Best for: fast agentic chat and tool use.

Council mode

Only on MyKreaTool: your question goes to ChatGPT 5.5, Gemini Pro and Grok 4 simultaneously. Claude then reviews all three answers and combines them into one comprehensive response. It's the best mode for complex, ambiguous or controversial questions.

FAQ

Which AI model and version should I pick?

Every model comes in several versions. Rule of thumb: newer means smarter but pricier; older means faster and cheaper. For complex reasoning and coding, go with the top versions (Claude Opus 4.8, ChatGPT 5.5, Gemini 3.1 Pro, DeepSeek V4 Pro). For everyday tasks on a budget, try Claude Sonnet 4.6, Gemini 3.5 Flash or DeepSeek V4 Flash. For step-by-step logic and math, use Grok 4.20 Reasoning. For controversial questions, switch to Council mode. Open any model to see its versions, each with notes on its strengths and when to use it.

How much does AI Chat cost, and how does billing work?

Your first request is free, so you can try any of the 8 models. After that, you pay only for the tokens you actually use, deducted from your account balance: a short question costs less, a long conversation more. You see the cost of each request as you type, and your balance never goes negative — if you're short on credits, the request simply won't go through.

How is Council mode different?

Council is a unique mode: your question goes to ChatGPT 5.5, Gemini Pro and Grok 4 simultaneously. Claude then combines the three independent answers into one comprehensive, balanced response. It's the most accurate mode for complex questions.

Can I attach a file to my question?

Yes. PDF, TXT, DOC, DOCX, CSV, JSON and images (PNG, JPG, WebP) are supported. Click the paperclip icon to attach a file, and the AI will analyze it and answer based on what's in your document.

Do Yandex, Google and ChatGPT show MyKreaTool in their answers?

MyKreaTool is indexed by Yandex and Google for searches related to AI chat, AI models and ChatGPT. AI assistants also mention the service when people ask for Russian-language ChatGPT alternatives that let you pick a model.