42 lines
1.2 KiB
Markdown
42 lines
1.2 KiB
Markdown
---
|
|
layout: default
|
|
---
|
|
|
|
# ¿Cómo se Cuentan los Tokens?
|
|
|
|
<div class="mt-4 text-sm">
|
|
|
|
<v-click>
|
|
|
|
<div class="mb-4">
|
|
|
|
Los tokens se cuentan en **entrada (input)** y **salida (output)** por separado
|
|
|
|
</div>
|
|
|
|
</v-click>
|
|
|
|
<v-click>
|
|
|
|
<div class="border-l-4 border-purple-500 pl-4 bg-purple-50 dark:bg-purple-900/20 p-3 mb-4 text-xs">
|
|
<h3 class="text-sm font-bold mb-2">💰 Costos</h3>
|
|
<p class="mb-2">Los LLMs cobran por tokens procesados</p>
|
|
<p class="mb-2">Output suele ser <strong>más caro</strong> que input (3-5x)</p>
|
|
<p class="mb-2">Ejemplo: GPT-4o → $2.50/1M tokens input, $10/1M tokens output</p>
|
|
<p>📊 <a href="https://raw.githubusercontent.com/BerriAI/litellm/refs/heads/main/model_prices_and_context_window.json" target="_blank">Precios actualizados por modelo (JSON)</a></p>
|
|
</div>
|
|
|
|
</v-click>
|
|
|
|
<v-click>
|
|
|
|
<div class="border-l-4 border-blue-500 pl-4 bg-blue-50 dark:bg-blue-900/20 p-3 text-xs">
|
|
<h3 class="text-sm font-bold mb-2">🔧 Herramienta para Contar Tokens</h3>
|
|
<p class="mb-2">OpenAI Tokenizer (funciona para la mayoría de modelos):</p>
|
|
<p><a href="https://platform.openai.com/tokenizer" target="_blank">platform.openai.com/tokenizer</a></p>
|
|
</div>
|
|
|
|
</v-click>
|
|
|
|
</div>
|