Tracked inference demand hit 44.7 trillion tokens per week, pushing the Token Demand Index to 189.9 — up 90% since 2026-05-06 and +74% over the past thirty days. The narrative of collapsing token prices obscures the reality that the market is expanding rapidly, with open-source models now handling 45.1% of volume at a demand-weighted effective price of just $2.18 per 1 million tokens. It is the output-side companion to the Compute Tightness Index (the GPU input).
As of 2026-06-28, tracked AI inference runs at 44.7T tokens/week — up +74% in 30 days and +90% since 2026-05-06. Open-source carries 45.1% of those tokens, holding the demand-weighted price near $2.18/1M — a fraction of frontier list prices even as frontier models themselves got pricier.
⤓ Download full daily series (CSV)| Group | Blended $/1M | Tokens/week | Share | 30d demand |
|---|---|---|---|---|
| Open-source | $0.40 | 20.2T | 45.1% | ▲ +110% |
| Other | $0.93 | 10.8T | 24.3% | ▲ +202% |
| Claude | $10.00 | 6.5T | 14.5% | ▲ +18% |
| OpenAI | $3.47 | 3.2T | 7.1% | ▲ +17% |
| $0.85 | 3.8T | 8.6% | ▼ -8% | |
| xAI | $1.56 | 129.4B | 0.3% | ▲ +16% |
Open-source = aggregated weights-available labs (DeepSeek, Qwen, Llama, Mistral, etc.). "Other" = unclassified OpenRouter providers. Share = % of tracked tokens. *Implied spend = list price × tokens, not realized revenue.
Built from OpenRouter's model rankings + pricing, grouped into Claude / OpenAI / Google / xAI / Open-source / Other. Token counts are OpenRouter's trailing-7-day figures (the rankings "week" view) — i.e. each point is tokens processed over the prior week, not a single day. Per group: blended median $/1M (75/25 input/output) and weekly total_tokens. Demand index = latest weekly total ÷ the first week's total × 100. Effective $/1M = demand-weighted blended price (as-of prices carried forward across snapshots where only one side updated). Series since 2026-05-06; it grows as new snapshots land. Full history as CSV. Coverage is OpenRouter traffic — a large but partial slice of the market.
Pair it with the Compute Tightness Index (GPU input) for both sides of the AI-compute economy. Query the raw data from an agent via the Bargo MCP endpoint (get_inference_economics).