Model Registry
This document explains the structure of the model registry available on Lawthropic.
Registry Structure
Models are organized by provider with the following information:
Entry Format
- Model Name: Unique identifier of the model
- Usage Factor: Multiplier indicating how token consumption is calculated
Usage Factors
Usage factors determine how token consumption is calculated for each model:
- 2x Factor: Consumption is double the actual token count
- 1x Factor: Consumption equals actual token count
- 0.4x Factor: Consumption is 40% of actual token count
Supported Providers and Models
GPT
| Model | Usage Factor |
|---|---|
| gpt-6-astra (flagship) | 5x |
| gpt-5.6-sol | 2x |
| gpt-5.6-terra | 1x |
| gpt-5.6-luna | 0.4x |
| gpt-5.5 | 2x |
| gpt-5.4 | 1x |
| gpt-5.4-mini | 0.3x |
| gpt-image-2 | 1x |
Claude
| Model | Usage Factor |
|---|---|
| claude-fable-5-1 (flagship) | 9.5x |
| claude-opus-5 | 2x |
| claude-opus-4.8 | 2x |
| claude-opus-4.6 | 2x |
| claude-sonnet-5 | 0.8x |
| claude-sonnet-4.6 | 1.2x |
| claude-haiku-4.5 | 0.4x |
Gemini
| Model | Usage Factor |
|---|---|
| gemini-3.7-flash | 1.2x |
| gemini-3.6-flash | 1.2x |
| gemini-3.5-flash | 1.2x |
| gemini-3.1-pro | 1.2x |
| gemini-3-flash-preview | 1.2x |
Grok
| Model | Usage Factor |
|---|---|
| grok-4.6 | 1.15x |
| grok-4.5 | 1.15x |
Practical Usage
When sending requests to these models, the system automatically applies the corresponding usage factor to calculate your token balance deduction.
For example:
- Using
claude-opus-4.8deducts 2 tokens for every processed token (2x factor) - Using
claude-haiku-4.5deducts 0.4 tokens for every processed token (0.4x factor)
This enables precise token tracking to optimize your budget.