lanista
a #B4mad project
model catalog & picker · 2,700+ models

The right model
for your task

A lanista owned a gladiator troupe: he knew each fighter's strengths, weaknesses, and price, and chose which to field for a given match. This tool does the same for LLMs.

lanista aggregates live data from eight catalog sources — OpenRouter, LiteLLM, LMArena, and more — plus practitioner opinions from three community feeds. Every recommendation is citation-grounded: each claim cites a specific catalog column or opinion ID.

lanista builds a self-contained prompt from the top-60 models by benchmark rating, curated tier notes, and 40 recent practitioner opinion excerpts — ready to paste into any LLM for a cited answer.

  • pareto — pure math: models not dominated on a quality/cost frontier.
  • profiles — collapses the frontier to three picks: flagship, balanced, budget.