I got tired of juggling OpenRouter + Artificial Analysis + Design Arena tabs to pick a model, so I put them in one filterable table

So every time I pick a model for a feature or random use-case I have I end up having like 12 tabs open — usually OpenRouter for price and context, Artificial Analysis for benchmarks, Design Arena for the UI/frontend Elo if thats relevant, a status/model page for throughput or other details. Got very fed up very quick so I built one table that joins all of it. [modelgrep.com](http://modelgrep.com) pulls \~300 models from OpenRouter live and lets you filter by: \- intelligence / coding / agentic index (Artificial Analysis) \- Design Arena Elo (human head-to-head for UI & frontend output) \- live throughput + time-to-first-token \- price, context length, vision/tools/reasoning/JSON support \- free API to pull all the same data if you need it somewhere So you search for stuff like "smartest model under $1/M with 200k+ context" or "fastest model with vision" in one go. Obviously free, no signup or no API key. Btw benchmark coverage is kinda uneven (not every model is scored), and "best for X" is dependent on the underlying index (which is pretty comprehensive but not perfect) Mostly looking for feedback here: what filters/intents would you actually use? Is the Design Arena angle useful? How can I make the UI/UX better for y'all? Tbh anything you have on your mind Repo is also opensource if you wanna run locally or mess around with it: [https://github.com/sculptdotfun/modelgrep](https://github.com/sculptdotfun/modelgrep)