LenSA model pricing

The strongest model in the lineup: architecture, deep analysis and work where a mistake costs more than the tokens.

Input

6,000≈ $15

Output

33,000≈ $82

Cache read

250≈ $1

Cache write

11,000≈ $27

OpenAI's flagship: long chains of reasoning, untangling contradictory requirements, mission-critical code.

Input

6,000≈ $15

Output

33,000≈ $82

Cache read

700≈ $2

Cache write

9,000≈ $22

Large refactorings, unfamiliar codebases and thorough review — where careful edits matter most.

Input

2,500≈ $6

Output

14,000≈ $35

Cache read

150

Cache write

3,800≈ $9

Meta's first frontier model for demanding multi-step and agentic work: it keeps long chains of actions on track and recovers from errors. Works with documents.

Input

2,500≈ $6

Output

15,000≈ $37

Cache read

150

Cache write

3,500≈ $9

The everyday all-rounder: complex logic, large edits and tidy code.

Input

1,300≈ $3

Output

7,000≈ $17

Cache read

100

Cache write

1,800≈ $4

An upgraded Kimi K3 from Moonshot, strong at complex coding and agentic terminal work. Reasoning effort is configurable.

Input

1,300≈ $3

Output

7,000≈ $17

Cache read

150

Cache write

2,000≈ $5

Google's new Gemini 4 flagship for complex coding, deep analysis and multi-step work. Works with documents; reasoning effort is configurable.

Input

1,200≈ $3

Output

6,000≈ $15

Cache read

100

Cache write

150

A fast model for unconventional solutions.

Input

1,200≈ $3

Output

4,000≈ $10

Cache read

350≈ $1

Cache write

2,000≈ $5

A strong OpenAI model for complex coding, deep analysis and multi-step work. Delivers near-Astra capability at a lower cost, with configurable reasoning effort.

Input

1,000≈ $2

Output

6,000≈ $15

Cache read

90

Cache write

1,700≈ $4

A model for coding and complex tasks.

Input

700≈ $2

Output

3,000≈ $7

Cache read

100

Cache write

1,200≈ $3

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

Input

700≈ $2

Output

3,500≈ $9

Cache read

50

Cache write

1,000≈ $2

Economical work on queries, code and algorithms.

Input

400≈ $1

Output

1,400≈ $3

Cache read

20

Cache write

700≈ $2

A high-speed reasoning assistant: fast refactoring, complex query generation, and 1C/BSL coding with massive context support.

Input

400≈ $1

Output

2,700≈ $7

Cache read

100

Cache write

100

A faster, less capable variant of GLM 5.3.

Input

200

Output

1,000≈ $2

Cache read

80

Cache write

400≈ $1

The cheapest option: simple transformations, drafts and bulk minor edits.

Input

100

Output

250≈ $1

Cache read

50

Cache write

180

A fast, inexpensive MiniMax model for everyday work: coding, tools and reasoning with configurable depth. Its answers can be verbose.

Input

100

Output

500≈ $1

Cache read

10

Cache write

150

An inexpensive Xiaomi model for reasoning, coding and frequent tool calls. Its large context window helps process lengthy material.

Input

100

Output

250≈ $1

Cache read

10

Cache write

150

The best value in the lineup: analysis, code and complex tasks at a fraction of the flagship price.

Input

100

Output

2,000≈ $5

Cache read

80

Cache write

100

A fast assistant for questions, short texts and code. Good for an active back-and-forth: explain, rephrase, suggest options and quickly refine the result.

Input

80

Output

400≈ $1

Cache read

10

Cache write

100

A very fast, inexpensive Inception diffusion model that answers almost instantly. Supports reasoning and tools; handy for frequent requests and code edits.

Input

80

Output

250≈ $1

Cache read

10

Cache write

110

A cost-efficient OpenAI model for focused tasks: small code edits, data extraction, short summaries and recurring automation. Supports configurable reasoning effort.

Input

80

Output

400≈ $1

Cache read

10

Cache write

100