Token Pricing

Token Pricing

Purchase Tokens

Tokens are used to run AI-powered automated analytical processes on theCodeSherpa. Tokens are stored in your Token Bank and do not expire.

3,000,000 Tokens
$25
7,500,000 Tokens
$50
18,000,000 Tokens
$100
45,000,000 Tokens
$200
Best Current Value
Log In to Purchase Tokens

Payments are processed by Jubilee Auto Rentals, LLC and may appear under that name on your payment records.

What Determines Token Usage?

Token usage depends on several factors, including the selected AI large language model, the size and complexity of your project, the number of Workers and functions being analyzed, and the amount of engineering context required to generate the Prompt Package.

Available AI Models

CodeSherpa currently provides three GPT-5.6 model options for Agentic Codebase Search, Prompt Generation, and other supported AI analytical processes:

Model Token-Cost Multipliers

CodeSherpa converts the underlying AI processing cost into CodeSherpa Tokens. More capable models have higher processing costs and therefore consume Tokens at higher rates.

AI Model Token-Cost Multiplier Best Suited For
GPT-5.6 Luna Base Token Cost × 3 Routine function analysis, focused requests, and general CodeSherpa use
GPT-5.6 Terra Base Token Cost × 12 Cross-Worker requests, complex dependencies, and substantial engineering analysis
GPT-5.6 Sol Base Token Cost × 50 Large architecture requests, high-complexity systems, and maximum-depth analysis

Choosing the Right Model

GPT-5.6 Luna is the recommended starting point for most CodeSherpa requests. It provides strong engineering analysis while using substantially fewer Tokens than Terra or Sol.

GPT-5.6 Terra is appropriate when a request spans multiple Workers, involves substantial architecture, or requires deeper dependency reasoning. GPT-5.6 Sol should generally be reserved for the most complex and consequential engineering requests, where maximum analytical depth is more important than Token efficiency.

Illustrative Prompt Package Token Usage

Actual usage varies considerably, but a substantial Agentic Prompt Package or Codebase Search may produce charges similar to the following:

AI Model Illustrative Token Usage
GPT-5.6 Luna Approximately 350,000 to 450,000 Tokens
GPT-5.6 Terra Approximately 1,400,000 to 1,700,000 Tokens
GPT-5.6 Sol Approximately 6,000,000 to 7,000,000 Tokens

These examples are illustrations based on substantial CodeSherpa analytical requests, not fixed prices. Actual Token usage can be much lower or higher depending on the request type, project size, selected Workers, number and size of functions, schemas, architecture context, model response length, and complexity of the analysis.

Our General Model Recommendation

For most Prompt Generator and Agentic Codebase Search requests, we recommend GPT-5.6 Luna. It offers a strong balance of analytical quality, speed, and Token efficiency.

Move to GPT-5.6 Terra when the request requires deeper cross-Worker or architectural reasoning. Use GPT-5.6 Sol selectively for exceptionally complex work where the additional analytical depth is likely to materially improve the result.

Current Purchase-Value Reference

The current Token packages provide progressively better value: 3,000,000 Tokens for $25; 7,500,000 Tokens for $50; 18,000,000 Tokens for $100; and 45,000,000 Tokens for $200. Because actual AI usage varies, CodeSherpa displays estimated charges before supported analytical processes are submitted.

Tokens Never Expire

Tokens placed in your Token Bank do not expire. They remain available until you use them or, after the Token Marketplace launches, sell or transfer eligible excess Tokens.

Token Marketplace

The Token Marketplace is scheduled for launch on October 1, 2026. The planned Marketplace will allow eligible Users to sell or transfer Tokens, subject to the Marketplace rules and fees published at launch.

The Marketplace is also planned to support the sale of products and services. This may include reusable software elements, templates, modules, and other resources that developers create for use in assembling applications.

Marketplace features, launch timing, eligibility requirements, pricing, transaction fees, and supported products may change before public release.