Skip to content
AI HOT (Curated Pool)

OpenRouter launches new Auto router that picks models based on community usage

OpenRouter 推出由市场智慧驱动的新版 Auto 路由器

OpenRouter turned its 55T weekly token usage data into a routing strategy. The new Auto router picks models based on what the community actually used for similar tasks over the past 7 days, generating a Pareto-optimal routing curve. At the default tier, MMLU Pro hits 85.2% while cost drops from $393 to $141. At max tier, SWE-Atlas QnA jumps from 2.4% to 60.7% but cost rises to $1,325. The router adds sticky behavior to avoid switching models mid-conversation and rebuilding caches. The post doesn't spell out how task classification works or how often the routing curve refreshes.

Why it matters: OpenRouter built its new Auto routing policy from 55 trillion tokens of real weekly community usage, scoring 85.2% on MMLU Pro at the default tier while cutting cost from $393 to $141.

Read the original ↗Export Markdown