Model Picker

The model menu of an AI chat, with the cost of the next reply for this chat.

PROGet Model Picker and 168 other Pro components with pro lifetime — Get Pro

Category
AI Agent
Access
Pro
Source
661 lines · one file
States
5 drawn
Added
21 Oct 2026

The decision inside it

Price for this conversation, not per million tokens

"$3 / M input" means nothing to the person choosing. Pass the size of the chat and the tray says "≈ $0.04 for the next reply". A model whose context window cannot hold the chat is disabled and says so in numbers.

When to use it

Use Model Picker in a chat composer or agent settings to choose a model with its speed, quality and the cost of the next reply for this conversation. Pass contextTokens so models that cannot hold the chat are disabled.

Installation

How you get it

Paste this into Claude Code, Cursor, Codex or any coding agent. It installs the file and uses it as it is, instead of writing a new one.

Use Liro's Model Picker for this.
Install it: npx shadcn@latest add @liro/model-picker
(the @liro registry is set up in components.json, with LIRO_KEY in .env.local)
Import it from @/components/ui/model-picker.
Do not rewrite or restyle it. Pass my data through its props
and keep its built-in states.

For the full brief with every prop and the mistakes assistants usually make, use Copy prompt above. Pro components install through the @liro registry with your CLI token: see the Namespaced tab.

Usage

components/chat-model.tsx
"use client";

import * as React from "react";
import { ModelPicker, type ModelOption } from "@/components/ui/model-picker";

const models: ModelOption[] = [
  { id: "swift", name: "Swift", provider: "Madgixs AI", speed: 5, quality: 3, inputPrice: 0.8, outputPrice: 4, contextWindow: 200_000, firstTokenMs: 300, recommended: true },
  { id: "deep", name: "Deep", provider: "Madgixs AI", speed: 2, quality: 5, inputPrice: 15, outputPrice: 75, contextWindow: 200_000, firstTokenMs: 1200, thinking: true },
];

export function ChatModel({ tokens }: { tokens: number }) {
  const [model, setModel] = React.useState("swift");
  return <ModelPicker models={models} value={model} onChange={setModel} contextTokens={tokens} side="top" />;
}

API reference

“Copy prompt” above puts a full brief on your clipboard, install command, dependencies, props, and the mistakes an assistant usually makes with this component.

Props

Read from the component’s own TypeScript interface when the site is built, so this table cannot drift from the file you install.

TypeNotes
modelsModelOption[] Yes–
valuestring NoThe chosen model id. Leave out to let the picker keep its own.
defaultValuestring No–
onChange(id: string) => void No–
thinkingThinkingState NoControlled thinking.
defaultThinkingThinkingState No–
onThinkingChange(t: ThinkingState) => void No–
contextTokensnumber NoTokens in the conversation so far. Drives the estimate and which models fit.
expectedOutputTokensnumber NoTokens a typical reply runs to. Default 800.
thinkingBudgetRecord<Effort, number> NoExtra output tokens per effort level. Default 1k / 4k / 16k.
openboolean NoControlled open state.
defaultOpenboolean No–
onOpenChange(open: boolean) => void No–
side"top" | "bottom" NoWhere the menu opens. Default `bottom`.

Dependencies

npm packages

clsxframer-motionlucide-reacttailwind-merge

Liro foundations, installed alongside it

liro-theme

States

All of them are drawn by the component. None is left for you to design.

  • Closed
  • Open
  • Thinking on
  • Too small for this chat
  • Searching

Accessibility

  • Keyboard reachable throughout, in visual order
  • Focus trapped in overlays, Escape closes them
  • Focus returns to whatever opened it
  • Live regions announce what changed, not that something did
  • Honours prefers-reduced-motion

Responsive behavior

  • Sizes to its container, not the viewport
  • Usable from 320px up
  • Pointer-based interaction, so trackpad and touch both work
  • No hover-only affordances

Used in a screenAll screens

ScreenPro

Assistant Home

The start page of an AI assistant: a greeting, the prompt box, connected apps and suggestions from them.