Built-in AI models
AI-COMFORT Model Catalog
Compare the models available through AI-COMFORT by task and price. Open a card for the full description and rates; you can also bring compatible models through BYOK.
Current CBR rate: 1 USD = 84.93 ₽
A model inside a workspace
Choose by the task—not just the name
The team can use built-in models, owner-managed BYOK models and permitted personal connections. Knowledge, tools and access rights determine what a model can actually do in the workspace.
Two pricing layers
The plan pays for the workspace. Built-in models are charged for actual tokens or generated images. BYOK and personal connections are billed by their providers.
View workspace plansCatalog
Current model lineup
The prices below are for AI-COMFORT's built-in models. Open a card for its full description and exact dollar rates. Ruble estimates use the exchange rate shown above.
01
Text and work tasks
Compare the models by task, speed and token price. Full descriptions and cached-input rates are inside each card.
02
Image generation
These models are billed per image; the amount shown is the catalog rate. Model-specific options and restrictions are described inside.
Pricing and access
Details, when you need them
The catalog stays easy to compare; the billing and configuration rules remain available here.
01Input, output and tokensWhy one request has more than one rate
A text request includes input tokens—the prompt and context sent to the model—and output tokens generated in its answer. Corporate instructions also count as input. Token counts can differ between models because they split text differently. Each model card shows its own rates.
02Save on repeated contextWhen repeated context costs less
Models with prompt caching can reuse context they have already processed, such as system settings, chat history, and documents, so these input tokens cost less. AI-COMFORT uses the provider-reported cached-token count and applies the reduced rate directly to the customer charge; the rate is shown on each supported model card.
03Minimum chargeOne kopeck per input and output component
The minimum charge is 0.01 ₽ for input and 0.01 ₽ for output separately. A calculated amount below one kopeck is rounded up for that component.
04Vision and imagesUnderstanding an image is not the same as generating one
A vision-capable text model can read and analyze an attached image, screenshot or document. Image-generation models create or edit visuals and are billed per image rather than per 1,000 text tokens.
05Corporate Model SettingsHow models work inside the team workspace
Built-in and BYOK-connected models do not remain raw APIs. AI-COMFORT adds a corporate system prompt, high reasoning, work-ready temperature, and available tools. This helps the model understand the platform rules, retain requirements, relationships, and intermediate results during a long dialogue, and handle complex multi-step tasks more reliably.
06Usage and spending controlOrganization budget and individual quotas
Set a shared ruble spending limit for built-in models by calendar day, week, or month. Once the limit is reached, new requests to built-in models pause until the next period begins. An administrator can set a daily limit for a specific available model — built-in or connected through BYOK — for each employee. This helps assign more expensive models to the tasks that need them. The dashboard's shared spending limit applies to built-in models. BYOK models are billed directly by the connected provider, but individual employee limits can also be set for them.
Build your team's model set
Start with built-in models, then add your own connections where they fit your work.