Model catalog

The balanced everyday Yenisei model

Our model for development: code, refactoring, review

Scan recognition for PDFs without a text layer

An experimental Yenisei model at a nominal price

Our image model: a picture from a description

The flagship Yenisei model for complex tasks and code

Speech recognition: voice input in the chat and the app

The all-purpose DeepSeek chat model

Fast DeepSeek V4 for high-volume tasks

Fast DeepSeek that sees images

The flagship DeepSeek V4: reasoning and code

Kimi K2 with a thinking mode

A recent Kimi with reasoning and agents

Kimi for code and development agents

The new generation of Kimi with a huge context

Next-generation Qwen code model, low cost

The largest Qwen of the 3.8 line, with reasoning

The latest all-purpose Qwen with reasoning

A fast, lightweight Qwen

Xiaomi's MiMo model with reasoning

Flagship MiMo from Xiaomi with reasoning and code

Fast GLM with reasoning and agents

Top new-generation GLM with a huge context

Flagship GLM: very large context and reasoning

Huge context, images and video, at a low price

Cohere's top model for enterprise tasks

IBM's practical model for business tasks

The most affordable reasoning model in the catalog

The flagship Llama 4 for text and images

The light Llama 4: images and long context

Microsoft's compact model with strong logic

A compact, fast Mistral

The balanced Mistral Small

Mistral's audio model: it understands speech

NVIDIA's fast model with reasoning

Thinking Machines' model: text, images, audio

Open Gemma 4: reasoning, text and images

The larger open Gemma 4

OpenAI's large open model

OpenAI's light open model