Fix token pricing precision by changing attributes from int to float

[jfrench9]

Summary

Updates the TokenPricing model to use float data types instead of integers for input and output token pricing attributes, enabling more precise cost calculations.

Key Changes

• Modified TokenPricing model attribute types from int to float
• Enhanced precision for token cost calculations
• Improved accuracy in pricing computations

Impact

• Breaking Change: This change modifies the data model and may affect existing integrations that expect integer values
• Applications consuming the TokenPricing model will now receive float values instead of integers
• Database schema may require migration if persisting these values

Testing Considerations

• Verify that all token pricing calculations work correctly with float values
• Test backward compatibility with existing API consumers
• Validate precision improvements in cost computation scenarios
• Ensure proper handling of decimal values in pricing workflows

Technical Notes

• The change affects the core pricing model used throughout the robosystems client
• Consider updating any validation rules that may have assumed integer constraints
• Review serialization/deserialization logic to ensure proper float handling

---

🤖 Generated with Claude Code

Branch Info:

• Source: bugfix/token-float
• Target: main
• Type: bugfix

Co-Authored-By: Claude noreply@anthropic.com