Module cost 

Axis 6: cost (input+output tokens × per-model pricing).

Pricing is richer than just (input, output). Modern frontier models bill differently for:

• Cached input tokens: Anthropic prompt caching, OpenAI prompt caching — typically 10% of the uncached rate.
• Reasoning tokens: GPT-5+ reasoning / o1-style thinking tokens exposed via usage.completion_tokens_details.reasoning_tokens — usually billed at the output rate.
• Batch API: OpenAI and Anthropic batch APIs are ~50% off.

Unknown models still contribute 0 cost rather than phantom infinities.

Structs

ModelPricing

Per-model pricing in USD per token.

Functions

compute

Compute the cost axis.

Type Aliases

Pricing

Pricing table keyed by chat_response.payload.model.