> ## Documentation Index
> Fetch the complete documentation index at: /docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Anthropic

> Use Claude models from Anthropic in your BoostGPT AI teammates

## Overview

Anthropic's Claude models are known for their safety, nuance, and strong performance on complex tasks. Claude excels at detailed analysis, creative writing, and maintaining helpful, harmless, and honest conversations.

## Available Models

> [card] Claude Fable 5.1
>
> **35 credits** • Anthropic's most capable model, with cheaper cache reads than Fable 5 for long agentic work.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Slow • Cost: High
> * Vision: images supported

> [card] Claude Opus 5.5
>
> **15 credits** • Anthropic's recommended model for most work: long-running agentic coding and knowledge work, near Fable 5.1 at a lower cost.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-5-5`

> [card] Claude Fable 5
>
> **39 credits** • Anthropic's most capable model, for the most demanding reasoning and long-running agentic work.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Slow • Cost: High
> * Vision: images supported
> * Model id: `claude-fable-5`

> [card] Claude Opus 5
>
> **20 credits** • Previous-generation Opus model for complex agentic coding and enterprise work, with a 1M-token context window.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-5`

> [card] Claude Opus 4.8
>
> **20 credits** • The most capable Claude model, built for complex agentic coding and enterprise work.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-4-8`

> [card] Claude Sonnet 5.5
>
> **8 credits** • The best combination of speed and intelligence in the current Claude lineup.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Fast • Cost: Medium
> * Vision: images supported
> * Model id: `claude-sonnet-5-5`

> [card] Claude Sonnet 5
>
> **8 credits** • The best combination of speed and intelligence, with near-Opus quality on coding and agentic work.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Fast • Cost: Medium
> * Vision: images supported
> * Model id: `claude-sonnet-5`

> [card] Claude Opus 4.7
>
> **20 credits** • Highly autonomous Claude Opus model, strong on long-horizon agentic work, vision, and memory.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-4-7`

> [card] Claude Opus 4.6
>
> **20 credits** • Claude Opus model with superior reasoning, creativity, tool use, and contextual understanding.
>
> * 1M context window
> * Exceptional reasoning (reasoning model)
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-4-6`

> [card] Claude Sonnet 4.6
>
> **12 credits** • Fast Claude Sonnet model with strong reasoning, creativity, tool use, and contextual understanding.
>
> * 1M context window
> * Excellent reasoning (reasoning model)
> * Speed: Fast • Cost: Medium
> * Vision: images supported
> * Model id: `claude-sonnet-4-6`

> [card] Claude Haiku 4.5
>
> **4 credits** • The fastest Claude model, with near-frontier intelligence for everyday tasks.
>
> * 200K context window
> * Good reasoning
> * Speed: Fast • Cost: Low
> * Vision: images supported
> * Model id: `claude-haiku-4-5`

> [card] Claude Opus 4.5
>
> **20 credits** • Previous-generation Claude Opus with superior reasoning and contextual understanding.
>
> * 200K context window
> * Exceptional reasoning
> * Speed: Medium • Cost: High
> * Vision: images supported
> * Model id: `claude-opus-4-5`

## Setup

### Using BoostGPT-Hosted API Keys

**1. Select Claude Model**

In your BoostGPT dashboard, select any Claude model when creating or configuring your bot. No Anthropic API key needed!

**2. Choose Your Model**

Select based on your needs:

* **Claude Sonnet 5.5**: Best for most production use cases
* **Claude Opus 5.5**: Long-running agentic coding and knowledge work
* **Claude Fable 5.1**: Demanding reasoning and long-horizon agentic work
* **Claude Haiku 4.5**: Fast responses for high-volume apps

### Using Your Own Anthropic API Key

> [tab] Dashboard Setup
>
> **1. Navigate to Integrations**
>
> Go to [app.boostgpt.co](https://app.boostgpt.co/integrations) and select **Integrations**
>
> **2. Select Anthropic**
>
> Find and click on the **Anthropic** provider
>
> **3. Add API Key**
>
> Enter your Anthropic API key and select which AI teammates will use this key
>
> **4. Save Configuration**
>
> Click save to apply your custom API key
>
> > [!TIP]
> >
> > Using your own API key can reduce costs for high-volume applications and gives you direct control over rate limits.

> [tab] Core SDK
>
> ```javascript bot.js theme={null}
> import { BoostGPT } from 'boostgpt';
>
> const client = new BoostGPT({
>   project_id: process.env.BOOSTGPT_PROJECT_ID,
>   key: process.env.BOOSTGPT_API_KEY
> });
>
> // Create bot with Claude model
> const botResponse = await client.createBot({
>   name: 'My Claude Bot',
>   model: 'claude-sonnet-5-5',
>   instruction: 'You are a helpful and thoughtful assistant.',
>   max_reply_tokens: 1000,
>   status: 'active'
> });
>
> if (botResponse.err) {
>   console.error('Error:', botResponse.err);
> } else {
>   console.log('Bot created:', botResponse.response);
> }
> ```

> [tab] Router SDK
>
> ```javascript router.js theme={null}
> import { Router, TelegramAdapter } from '@boostgpt/router';
>
> const router = new Router({
>   apiKey: process.env.BOOSTGPT_API_KEY,
>   projectId: process.env.BOOSTGPT_PROJECT_ID,
>   defaultBotId: process.env.BOOSTGPT_BOT_ID, // Bot using Claude
>   adapters: [
>     new TelegramAdapter({
>       telegramToken: process.env.TELEGRAM_TOKEN
>     })
>   ]
> });
>
> // Router automatically uses your Claude bot
> router.onMessage(async (message, context) => {
>   // Handle commands, return null for AI
>   if (message.content === '/help') {
>     return 'Ask me anything!';
>   }
>
>   return null; // Let Claude handle it
> });
>
> await router.start();
> ```

## Reasoning Modes

Configure how your Claude AI teammate thinks and responds:

> [accordion] Auto Mode (Recommended)
>
> **1-10x credits** • Automatically selects the best reasoning approach based on your question's complexity.
>
> Configure reasoning mode in the dashboard when creating/editing your AI teammate, or via API when creating custom bots.
>
> Best for: Most use cases, intelligent cost optimization

> [accordion] Quick Mode (Standard)
>
> **1x credits** • Straightforward answers without additional analysis. Most efficient option.
>
> Best for: Simple queries, FAQs, high-volume applications

> [accordion] Research & Tools Mode (Interactive)
>
> **Up to 10x credits** • Uses tools to gather information, perform calculations, or execute tasks. Most powerful option.
>
> Best for: Complex queries requiring external data or tools

> [accordion] Step-by-Step Mode (Stepwise)
>
> **Up to 2x credits** • Breaks problems into smaller steps with sources for each part, improving transparency.
>
> Best for: Educational content, troubleshooting, explanations

> [accordion] Deep Thinking Mode (React)
>
> **Up to 5x credits** • Cycles through thinking and reflecting to refine answers progressively.
>
> Best for: Complex analysis, strategic planning, research

> [!TIP]
>
> Reasoning modes are configured in the BoostGPT dashboard when creating or editing your AI teammate. They control how the model approaches problem-solving and can significantly impact credit usage.

## Model Selection Guide

### When to Use Each Model

> [accordion] Claude Sonnet 5.5 - Production Workhorse
>
> **Best for:**
>
> * Customer support and chatbots
> * Content generation and editing
> * Agentic coding
> * Balanced cost and performance
>
> **Cost:** about 8 credits per step

> [accordion] Claude Opus 5.5 - Long-Running Work
>
> **Best for:**
>
> * Long agentic coding sessions
> * Complex analysis and research
> * Knowledge work across large contexts
>
> **Cost:** about 15 credits per step

> [accordion] Claude Fable 5.1 - Maximum Intelligence
>
> **Best for:**
>
> * Demanding reasoning
> * Long-horizon agentic work
> * High-stakes analysis
>
> **Cost:** about 35 credits per step

> [accordion] Claude Haiku 4.5 - Speed & Volume
>
> **Best for:**
>
> * Quick responses
> * Simple queries
> * High-volume applications
>
> **Cost:** about 4 credits per step

## Troubleshooting

> [accordion] High costs with reasoning modes
>
> **Cause:** Research & Tools (10x) and Deep Thinking (5x) modes multiply credit costs
>
> **Solutions:**
>
> * Use Quick mode (standard) for simple queries
> * Use Auto mode to let Claude optimize automatically
> * Reserve deep modes for complex tasks only
> * Monitor credit usage in dashboard

> [accordion] Slow responses
>
> **Cause:** Deep Thinking mode or Extended Thinking models are slower
>
> **Solutions:**
>
> * Use Claude Haiku 4.5 for faster responses
> * Use Quick mode (standard reasoning)
> * Add "thinking..." indicators for user feedback
> * Reserve slow modes for async tasks

> [accordion] Context length errors
>
> **Rare:** Claude's 200K context handles most use cases
>
> **Solutions:**
>
> * Implement message history pruning
> * Summarize older messages
> * Use streaming for very long outputs

## Next Steps

> [card] [Model Comparison](/docs/providers/model-comparison.md)
>
> Compare Claude with other providers

> [card] [Reasoning Models](/docs/providers/reasoning-models.md)
>
> Deep dive into Extended Thinking

> [card] [Bring Your Own Key](/docs/providers/bring-your-own-keys.md)
>
> Use your own Anthropic API key

> [card] [SDK Reference](/docs/sdk/core/api-reference.md)
>
> Full API documentation


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.