D'Lemma model pricing

Anthropic's fastest model, for an active back-and-forth, short texts and code. Decides on its own when and how deeply to think before answering. A good fit when a quick reply matters.

Input

100

Output

350≈ $1

Cache read

10

Cache write

100

A general-purpose model for analysis, drafting materials and programming. Breaks a task into steps, works through code and refines a solution iteratively.

Input

1,300≈ $3

Output

7,000≈ $17

Cache read

80

Cache write

1,700≈ $4

For complex tasks where depth of analysis and attention to detail matter. Suited to comparing approaches, untangling ambiguous questions and working with intricate code.

Input

2,500≈ $6

Output

13,000≈ $32

Cache read

150

Cache write

3,500≈ $9

Anthropic's flagship model for complex multi-step work: programming, document analysis and research. Always reasons before answering. A good fit when depth and accuracy matter more than the price of a reply.

Input

4,000≈ $10

Output

28,000≈ $70

Cache read

150

Cache write

7,000≈ $17

An affordable OpenAI model of the GPT-6 generation for everyday tasks: questions, short texts, edits and discussion. Suited to an active back-and-forth where speed and a low price per reply matter.

Input

100

Output

350≈ $1

Cache read

10

Cache write

100

A fast and efficient Open AI model for visual reasoning, precise computer control, and generating clear, naturally sounding text.

Input

800≈ $2

Output

3,000≈ $7

Cache read

40

Cache write

1,000≈ $2

An OpenAI GPT-6-generation model for complex work: coding, document analysis and multi-step tasks. Stays coherent across long texts and tasks with many conditions. A step up from everyday models in depth of analysis.

Input

1,200≈ $3

Output

6,000≈ $15

Cache read

90

Cache write

1,700≈ $4

For the hardest tasks that need multi-step reasoning. Suited to deep analysis, designing solutions and technical questions with many interacting conditions.

Input

6,000≈ $15

Output

30,000≈ $74

Cache read

700≈ $2

Cache write

7,000≈ $17

A Google model for explanations, text work and programming. Helps get into a topic, write a short summary, discuss a solution and analyse an image.

Input

150

Output

500≈ $1

Cache read

20

Cache write

200

A fast general-purpose Google model for everyday work. Suited to analysing materials, preparing texts, explaining difficult topics and coding tasks.

Input

500≈ $1

Output

2,500≈ $6

Cache read

70

Cache write

100

An updated Google model with stronger coding and reasoning. Helps work through a multi-stage task, cross-check information and think a technical solution through.

Input

500≈ $1

Output

3,000≈ $7

Cache read

70

Cache write

100

Google's flagship model for complex, multi-step tasks. Suited to programming, analysing extensive material and developing detailed solutions that require comparing constraints and reasoning through several steps.

Input

1,200≈ $3

Output

6,000≈ $15

Cache read

100

Cache write

150

An updated Meta model for programming and tasks with many interrelated conditions. In the Contributor edition, prompts and responses may be used to improve Meta's products.

Input

70

Output

150

Cache read

2

Cache write

100

Meta's flagship model for programming, document analysis and complex work tasks. Helps plan a solution, make sense of interrelated requirements and work through a task step by step.

Input

2,500≈ $6

Output

13,000≈ $32

Cache read

150

Cache write

3,500≈ $9

A fast Inception model built on a diffusion architecture: it generates several parts of an answer at once, so replies come with low latency. Suited to an active back-and-forth, summaries, search and coding tasks.

Input

140

Output

500≈ $1

Cache read

12

Cache write

200

A model from the Land of the Rising Sun: Sakana AI's model trained for the Japanese language and Japanese business customs. Suited to texts and correspondence in Japanese, translation between Japanese and English, and coding. Works through hard tasks by step-by-step reasoning.

Input

1,000≈ $2

Output

4,000≈ $10

Cache read

200

Cache write

1,500≈ $4

A model from xAI, the company founded by Elon Musk, for studying topics in depth, analysing information and programming. Helps weigh arguments, consider different approaches and work through a complex task step by step.

Input

1,200≈ $3

Output

4,000≈ $10

Cache read

300≈ $1

Cache write

1,800≈ $4

A new-generation DeepSeek model with image understanding. Helps with logical and technical problems, code review and matching text against visual materials.

Input

100

Output

500≈ $1

Cache read

15

Cache write

200

A general-purpose Xiaomi model for text, images and code. Helps analyse lengthy material, explain a solution and work through tasks in several steps.

Input

100

Output

200

Cache read

4

Cache write

120

A fast MiniMax model for programming, tool use and multi-step tasks. Understands images. Suited to working through large materials and code at a low price per reply.

Input

100

Output

400≈ $1

Cache read

20

Cache write

150

A fast Z.ai model for text, images and code. Always reasons before answering. Suited to reviewing materials, explaining diagrams and step-by-step tasks when a quick reply matters.

Input

250≈ $1

Output

850≈ $2

Cache read

70

Cache write

300≈ $1

A powerful general-purpose Alibaba model for analysis, programming and image work. Helps make sense of complex materials, reconcile the conditions of a task and prepare a detailed solution.

Input

1,200≈ $3

Output

4,500≈ $11

Cache read

200

Cache write

2,000≈ $5

A Moonshot AI model for complex analysis and coding tasks. Understands images and reasons before answering. Suited to working through large materials and developing a solution step by step.

Input

1,200≈ $3

Output

6,000≈ $15

Cache read

150

Cache write

2,000≈ $5

For detailed text analysis, expanded answers and creative tasks. Suited when several conditions must be met, an idea developed or substantial Russian-language material prepared.

Input

3,800≈ $9

Output

4,100≈ $10

Cache read

400≈ $1

Cache write

4,500≈ $11

A compact Sber model for short questions, everyday conversation and simple tasks in Russian. Its small size lets it answer quickly. A good fit when a fast reply matters more than depth of analysis.

Input

110

Output

110

Cache read

110

Cache write

140

Sber's latest-generation model for work and study. Helps get into a topic, organise information, prepare a text or discuss a coding solution.

Input

4,200≈ $10

Output

4,300≈ $11

Cache read

500≈ $1

Cache write

5,500≈ $14

A Sber reasoning model that works through a task step by step before answering. Further trained on solving maths and programming problems. Suited to calculations, coding tasks and questions with several conditions in Russian.

Input

800≈ $2

Output

2,200≈ $5

Cache read

800≈ $2

Cache write

1,000≈ $2

A fast Yandex model for everyday tasks with texts and documents. Summarises, structures a text into points and finds the data you need in a file.

Input

700≈ $2

Output

1,500≈ $4

Cache read

200

Cache write

1,000≈ $2

Yandex's flagship text model, strongest in conversation. Suited to an assistant in a chat, answering questions and working with Russian-language texts.

Input

4,000≈ $10

Output

10,000≈ $25

Cache read

4,000≈ $10

Cache write

5,000≈ $12

Yandex's YandexGPT 5.1 model for Russian-language texts and documents. Rewrites a text in the required style and format, extracts data and picks out the key points.

Input

6,000≈ $15

Output

6,000≈ $15

Cache read

6,000≈ $15

Cache write

6,000≈ $15