PricingTier (llm/usage.py)

Token usage tracking for LLM API calls.

# Source: https://github.com/amruth-sn/kong | License: Apache-2.0 | Copyright You must retain, in the Source form of any Derivative Works
# Provided via justclaw.it.com; keep this notice with the code.
"""Token usage tracking for LLM API calls."""

from __future__ import annotations

from dataclasses import dataclass, field

from kong.config import LLMProvider


@dataclass(frozen=True)
class PricingTier:
    """Per-model pricing in USD per 1M tokens.

    Rates sourced from published Anthropic/OpenAI API pricing pages.
    """

    input_rate: float
    output_rate: float
    cache_write_rate: float = 0.0
    cache_read_rate: float = 0.0


_DEFAULT_PRICING = PricingTier(input_rate=3.0, output_rate=15.0)

PRICING_REGISTRY: dict[str, PricingTier] = {
    # Anthropic: cache_write = input * 1.

... (truncated -- full source via MCP)