Type in Any Task. Get the Best AI Model for It

Task Router launched a routing service that directs AI requests by task alias (e.g. taskrouter/<task-id>) to the model meeting a task's quality floor under a cheapest, balanced or most accurate priority, backed by 268 measured tasks. Each task is graded on a frozen dataset with a task-specific rubric and an LLM judge, then joined to published provider pricing with source and freshness shown, and the company states results are directional rather than a production SLA. Users whose tasks are not listed can submit their current model, prompt and real examples, and Task Router benchmarks other models against the incumbent and emails the results.

Route your requests by task alias e.g. taskrouter/<task-id . Calls go to the model that fits your priority cheapest, balanced or most accurate among those that clear the task’s quality floor. Each task shows exactly what to send and what you get back. Your own tasks are detected in your traffic or defined from samples. Don’t see your task? Tell us what you do today, the model and prompt you use, and a few real examples. We benchmark other models against your current one and email you the results. Meanwhile, keep calling your model through Task Router. Every comparison uses the same task-specific process. Results are directional rather than a production SLA, and each task shows its sample, rubric, judge, limitations, and pricing provenance. 01 Freeze a dataset We create and freeze a representative set of real task cases, so every model is graded on exactly the same work. 02 Define the rubric A task-specific rubric spells out what counts as correct — and what counts as a failure — before any model runs. 03 Run every tier Qualified candidates from all four tiers run on the frozen holdout and are scored by an LLM judge for comparable pass rates. 04 Join published pricing Measured token profiles meet the latest available provider prices, with source and freshness shown so every estimate can be checked. Task Router · 268 measured tasks Tasks we’ve measured, ready to route. Route your requests by task alias e.g. taskrouter/<task-id . Calls go to the model that fits your priority cheapest, balanced or most accurate among those that clear the task’s quality floor. Each task shows exactly what to send and what you get back. Your own tasks are detected in your traffic or defined from samples. Describe your task The assistant finds matching tasks and their evidence, or helps you request a benchmark. Conversation New chat Ask Task Router Enter to send · Shift + Enter for a new line What is Task Router?What's the cheapest model scoring at least 90% on FDD Item 19 Summarizer?For FDD Item 19 Summarizer, which model and provider give the lowest cost above 90% accuracy?How do routing policies and execution plans work?Can a small model handle Dietary Allergy Menu Screening as well as a frontier model?Which Neurometric models changed price in the last two weeks?How do I use Playground?Compare quality, cost, and latency for Access Request Agent. AllExtractionClassificationGenerationSummarizationAnalysisResearchTool use Given a customer contract, performance obligations, delivery events, and invoices, determine the proposed recognition schedule and flag contract-data conflicts. Given synthetic business emails, classify likely privilege status, identify attorney involvement and legal-purpose signals, and route uncertain cases for review. Given bank transactions and general-ledger entries, match reconciling items and explain outstanding deposits, fees, timing differences, and unexplained variance.