The auctionBeta

Scoring: price, latency, quality

The auction ranks effective cost, not sticker price. Latency and quality are measured by the router, not claimed by providers.

Owner
Protocol
Version
second-score/3
Verified
2026-07-13

Version second-score/3 ranks effective cost rather than sticker price. Expected request cost is multiplied by measured latency and service factors; the lowest score wins. Money remains integer USDC base units. Floating-point scoring chooses a provider but never touches a stored balance.

The score inputs

InputSourceEffect
expected costsigned input, cached-input, and output rates × forecast tokensthe raw monetary base of the score
TTFTrouter-measured time to first tokenpriority-weighted latency markup, capped against the fastest candidate
reliabilityrouter-observed completion successpenalizes fast-first-byte but incomplete service
trustverification score for provider and model classpenalizes adverse integrity evidence
throughputrouter-measured output ratepenalizes materially slower completion than the measured field

New or clean providers are not assigned an automatic quality bonus: every service factor is 1.0 when there is no adverse signal. Raw expected cost remains separate so the customer cap is priced on genuine cost rather than a quality penalty.

Customer priorities

cheapest, balanced, and fastest change the exchange rate between expected cost and measured TTFT. Reliability, trust, and throughput protections remain in the service factor. The same clearing and disclosure rules apply to every preset.

sessions change the cost viewFor a continuation turn, the cache-holding incumbent is scored at its locked tariff and cached-prefix cost; challengers carry real re-prefill cost. If the incumbent wins it bills the existing lock rather than receiving a fresh score-derived price.
As a customerA cheap quote can lose when measured service makes it more expensive in practice. Presets change your price/latency preference without hiding the service evidence.
As a providerPrice is only one route to a better score. Completing reliably, preserving trust, improving TTFT, and sustaining throughput all reduce your measured effective cost.