LLM Token Expenditure Index Family

The LLM Token Expenditure Index measures the effective expenditure level of the large-language-model market, expressed as a price-per-million-token benchmark. It is not a vendor price sheet, a model leaderboard, or a raw usage chart. It is a daily market reference that combines model pricing with market activity so users can understand what it effectively costs to access the active LLM market. Within the index family, the broad-market index is complemented by Open LLM and Proprietary LLM views, allowing users to compare the active market with open-source/open-weight and closed-source model segments.

“Token price” is too narrow for the way LLM markets work. A customer does not experience one universal token price. They experience a mix of input tokens, output tokens, model choices, provider prices, and workload patterns. Two models can have similar posted prices but very different expenditure impact if one receives more market activity or has a different input-output mix. This is why the same framework can support the broad-market, Open LLM, and Proprietary LLM views: each view measures expenditure conditions for a defined model universe rather than a single posted token price.

The index is designed to follow sustained economic relevance, not catalog availability. Models enter the benchmark basket because they show meaningful market activity, not simply because they appear on a provider page. Older models can lose influence as usage fades, and emerging models can gain influence as market adoption becomes durable. The same principle applies across the Open LLM and Proprietary LLM series, where each segment follows sustained activity inside its own model universe.

The index can rise when providers increase prices, when usage shifts toward more expensive models, when output-heavy workloads become more important, or when high-expenditure models gain benchmark weight. It can fall when prices decline, when lower-cost models gain adoption, when users substitute toward more efficient model choices, or when the market diversifies away from expensive concentration. Across the index family, a move in the Open LLM view may point to changing open-model competitiveness, while a move in the Proprietary LLM view may reflect shifts in the premium paid for closed-model capability, integration, or distribution.

Customers can use the index as a market baseline for LLM cost conditions. Procurement teams can compare internal token spend against the broader market and separate vendor-specific exposure from market-wide movement. Finance teams can use it to monitor whether LLM access is becoming structurally more or less expensive. Product and platform teams can use it to understand whether their model-routing decisions are moving with or against the market. The index family also helps customers compare whether market movement is being led by open models, proprietary models, or neither. If the broad-market index begins to follow the Open LLM view lower, that may suggest open models are gaining enough economic weight to influence the market benchmark. If it follows the Proprietary LLM view instead, demand may still be set by closed-model capability and willingness to pay. If neither view explains the broad-market move cleanly, the change may reflect broader mix, composition, or input-output dynamics across the market.

The index does not disclose proprietary raw usage inputs, or private commercial terms. It is also not a guarantee of any customer’s realized invoice, because individual bills can depend on private discounts, routing policy, workload design, caching, context length, and provider-specific terms. The index is meant to be a high-integrity benchmark, not a replacement for account-level cost accounting. It is supported by validation checks around market coverage, model activity, basket composition, weight concentration, and day-to-day movement.