docs/MODELS.md

Models Guide

Reference for all Rox AI models.

Model Overview

Rox AI offers eight specialized models, each optimized for different use cases:

Model	Endpoint	Best For	Max Tokens	Default Temp
Rox Core	/chat	General conversation	4,096	1.0
Rox 2.1 Turbo	/turbo	Fast responses	4,096	0.6
Rox 3.5 Coder	/coder	Code generation	16,384	0.6
Rox 4.5 Turbo	/turbo45	Advanced reasoning	8,192	0.2
Rox 5 Ultra	/ultra	Superior reasoning	8,192	1.0
Rox 6 Dyno	/dyno	Extended context	16,384	1.0
Rox 7 Coder	/coder7	Advanced coding	16,384	1.0
Rox Vision Max	/vision	Visual understanding	512	0.2

---

Rox Core

Endpoint: POST /chat

Description

General-purpose conversational model for everyday tasks.

Best Use Cases

• General conversation and Q&A
• Content writing and generation
• Creative tasks (stories, poems, ideas)
• Summarization and analysis
• Educational tutoring
• Customer support

Parameters

• Temperature: 1.0 (balanced creativity)
• Top P: 1.0 (full diversity)
• Max Tokens: 4,096

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/chat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Explain quantum computing in simple terms"}
    ],
    "temperature": 1.0,
    "max_tokens": 512
  }'

When to Choose Rox Core

• You need creative, varied responses
• Task requires nuanced understanding
• Building a general-purpose chatbot
• Content needs to be engaging and natural

---

Rox 2.1 Turbo

Endpoint: POST /turbo

Description

Fast model for quick responses and real-time applications.

Best Use Cases

• Real-time chat applications
• Customer support bots
• Quick Q&A systems
• High-throughput applications
• Simple queries and commands
• Factual information retrieval

Parameters

• Temperature: 0.6 (more focused)
• Top P: 0.7 (more consistent)
• Max Tokens: 4,096

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/turbo \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "What are the business hours?"}
    ]
  }'

When to Choose Rox 2.1 Turbo

• Speed is critical
• Need consistent, reliable answers
• Building customer support systems
• High volume of requests
• Simple, straightforward queries

---

Rox 3.5 Coder

Endpoint: POST /coder

Description

Code-focused model for programming tasks and technical work.

Best Use Cases

• Code generation and completion
• Debugging and error fixing
• Algorithm design and optimization
• Technical documentation
• Code review and suggestions
• Software architecture discussions
• API integration help

Parameters

• Temperature: 0.6 (precise and focused)
• Top P: 0.95 (balanced diversity)
• Max Tokens: 16,384 (extended context)
• Special Features: Enhanced thinking mode

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/coder \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Write a Python function to implement binary search"}
    ],
    "max_tokens": 2048
  }'

When to Choose Rox 3.5 Coder

• Working with code in any language
• Need detailed technical explanations
• Debugging complex issues
• Designing algorithms or systems
• Writing technical documentation
• Need extended context (up to 16K tokens)

---

Comparison Matrix

Performance Characteristics

Feature	Rox Core	Rox 2.1 Turbo	Rox 3.5 Coder	Rox 4.5 Turbo	Rox 5 Ultra	Rox 6 Dyno	Rox 7 Coder	Rox Vision
Speed	Medium	Fast	Medium	Fast	Medium	Medium	Medium	Fast
Creativity	High	Medium	Low	Low	High	High	Medium	Low
Consistency	Medium	High	High	Very High	High	Medium	High	Very High
Code Quality	Good	Good	Excellent	Good	Excellent	Good	Superior	N/A
Context Length	4K	4K	16K	8K	8K	16K	16K	512
Thinking Mode	No	No	Yes	Yes	Yes	Yes	Yes	No
Reasoning	Basic	Basic	Advanced	Very Advanced	Superior	Advanced	Superior	Basic

Use Case Recommendations

Task	Recommended Model	Why
Write a blog post	Rox Core	Creative, engaging content
Answer "What is X?"	Rox 2.1 Turbo	Fast, factual response
Debug Python code	Rox 3.5 Coder	Code specialist
Customer support	Rox 2.1 Turbo	Quick, consistent answers
Write a story	Rox Core	Creative and varied
Explain algorithm	Rox 3.5 Coder	Technical depth
Translate text	Rox 2.1 Turbo	Fast and accurate
Design API	Rox 3.5 Coder	Technical expertise
Brainstorm ideas	Rox Core	Creative thinking
Code review	Rox 3.5 Coder	Code understanding
Complex reasoning	Rox 4.5 Turbo	Advanced thinking
Research analysis	Rox 5 Ultra	Superior reasoning
System architecture	Rox 5 Ultra	Complex design
Long documents	Rox 6 Dyno	Extended context
Large codebase	Rox 7 Coder	Advanced coding
Image analysis	Rox Vision Max	Visual understanding

---

Model Selection Guide

Decision Tree

Need to work with code?
├─ Yes
│  ├─ Simple/medium tasks? → Rox 3.5 Coder
│  └─ Complex/large-scale? → Rox 7 Coder
└─ No
   ├─ Need advanced reasoning?
   │  ├─ Yes
   │  │  ├─ Need highest quality? → Rox 5 Ultra
   │  │  └─ Need speed? → Rox 4.5 Turbo
   │  └─ No
   │     ├─ Long documents? → Rox 6 Dyno
   │     ├─ Visual tasks? → Rox Vision Max
   │     ├─ Need fast responses? → Rox 2.1 Turbo
   │     └─ Need creative output? → Rox Core

Quick Selection Tips

Choose Rox Core when:

• Default choice for most tasks
• Need creative, engaging responses
• Building general chatbots
• Content generation projects

Choose Rox 2.1 Turbo when:

• Speed matters most
• Need consistent answers
• High request volume
• Simple Q&A systems

Choose Rox 3.5 Coder when:

• Any coding task
• Technical documentation
• Algorithm design
• Need extended context

Choose Rox 6 Dyno when:

• Processing long documents
• Extended context needed
• Multi-document analysis
• Long conversations

Choose Rox 7 Coder when:

• Most complex coding tasks
• Large-scale projects
• System architecture
• Advanced algorithms

Choose Rox Vision Max when:

• Visual understanding
• Image analysis
• Multimodal tasks

---

Advanced Usage

Switching Models Dynamically

class RoxAI {
  constructor(baseUrl = 'https://Rox-Turbo-API.hf.space') {
    this.baseUrl = baseUrl;
  }

  async chat(message, model = 'chat') {
    const endpoints = {
      core: 'chat',
      turbo: 'turbo',
      coder: 'coder'
    };

    const endpoint = endpoints[model] || model;

    const response = await fetch(`${this.baseUrl}/${endpoint}`, {
      method: 'POST',
      headers: { 'Content-Type': 'application/json' },
      body: JSON.stringify({
        messages: [{ role: 'user', content: message }]
      })
    });

    return (await response.json()).content;
  }
}

// Usage
const rox = new RoxAI();

// Use different models for different tasks
const story = await rox.chat('Write a short story', 'core');
const answer = await rox.chat('What is 2+2?', 'turbo');
const code = await rox.chat('Write a sorting function', 'coder');

Model-Specific Optimization

import requests

class RoxClient:
    def __init__(self, base_url="https://Rox-Turbo-API.hf.space"):
        self.base_url = base_url
    
    def ask_core(self, message, creative=True):
        """Use Rox Core with creativity control"""
        return self._request('chat', message, 
                           temperature=1.2 if creative else 0.8)
    
    def ask_turbo(self, message):
        """Use Rox Turbo for fast responses"""
        return self._request('turbo', message, max_tokens=256)
    
    def ask_coder(self, message, extended=False):
        """Use Rox Coder with optional extended context"""
        return self._request('coder', message,
                           max_tokens=8192 if extended else 2048)
    
    def _request(self, endpoint, message, **kwargs):
        response = requests.post(
            f"{self.base_url}/{endpoint}",
            json={
                "messages": [{"role": "user", "content": message}],
                **kwargs
            }
        )
        return response.json()["content"]

---

Cost and Performance Optimization

Tips for Each Model

Rox Core:

• Use for tasks requiring creativity
• Adjust temperature based on needs
• Consider caching common queries

Rox 2.1 Turbo:

• Best cost-performance ratio
• Use for high-volume applications
• Lower max_tokens for even faster responses

Rox 3.5 Coder:

• Use only for code-related tasks
• Leverage extended context when needed
• Cache code snippets and patterns

---

API Compatibility

All three models use the same request/response format:

Request:

{
  "messages": [
    {"role": "user", "content": "Your message"}
  ],
  "temperature": 1.0,
  "top_p": 0.95,
  "max_tokens": 512
}

Response:

{
  "content": "Model response"
}

This makes it easy to switch between models without changing your code!

---

---

---

Built by Mohammad Faiz

Rox 4.5 Turbo

Endpoint: POST /turbo45

Description

Reasoning model for complex problem-solving with fast responses.

Best Use Cases

• Complex problem solving
• Advanced reasoning tasks
• Scientific explanations
• Mathematical problems
• Strategic planning
• Analysis and insights

Parameters

• Temperature: 0.2 (highly focused)
• Top P: 0.7 (consistent)
• Max Tokens: 8,192
• Special Features: Enhanced reasoning mode

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/turbo45 \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Explain the theory of relativity"}
    ],
    "max_tokens": 2048
  }'

When to Choose Rox 4.5 Turbo

• Need advanced reasoning
• Complex problem solving
• Scientific or technical explanations
• Fast responses with deep thinking

---

Rox 5 Ultra

Endpoint: POST /ultra

Description

Advanced model for complex reasoning and high-quality output.

Best Use Cases

• Most complex problem solving
• Research and analysis
• Advanced technical tasks
• Strategic decision making
• Complex code architecture
• Multi-step reasoning

Parameters

• Temperature: 1.0 (balanced)
• Top P: 0.95 (high diversity)
• Max Tokens: 8,192
• Special Features: Superior reasoning mode

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/ultra \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Design a scalable microservices architecture"}
    ],
    "max_tokens": 4096
  }'

When to Choose Rox 5 Ultra

• Most complex tasks
• Need highest quality output
• Multi-step reasoning required
• Research and deep analysis

---

Rox 6 Dyno

Endpoint: POST /dyno

Description

Extended context model for long documents and conversations.

Best Use Cases

• Long document analysis
• Extended conversations
• Document summarization
• Research paper analysis
• Multi-document synthesis

Parameters

• Temperature: 1.0 (balanced)
• Top P: 1.0 (full diversity)
• Max Tokens: 16,384 (extended context)
• Special Features: Dynamic thinking mode

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/dyno \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Analyze this 20-page document..."}
    ],
    "max_tokens": 8192
  }'

When to Choose Rox 6 Dyno

• Processing long documents
• Need extended context window
• Multi-document analysis
• Long-form content generation

---

Rox 7 Coder

Endpoint: POST /coder7

Description

Advanced coding model for complex programming tasks.

Best Use Cases

• Complex algorithm design
• Large-scale code generation
• Advanced debugging
• System architecture
• Code refactoring
• Multi-file code analysis

Parameters

• Temperature: 1.0 (balanced)
• Top P: 1.0 (full diversity)
• Max Tokens: 16,384 (extended context)
• Special Features: Advanced thinking mode for code

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/coder7 \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Build a distributed caching system"}
    ],
    "max_tokens": 8192
  }'

When to Choose Rox 7 Coder

• Most complex coding tasks
• Large-scale projects
• System design and architecture
• Advanced algorithms

---

Rox Vision Max

Endpoint: POST /vision

Description

Visual model for image analysis and multimodal tasks.

Best Use Cases

• Image analysis
• Visual understanding
• Multimodal tasks
• Image description
• Visual Q&A

Parameters

• Temperature: 0.2 (highly focused)
• Top P: 0.7 (consistent)
• Max Tokens: 512

Example Request

curl -X POST https://Rox-Turbo-API.hf.space/vision \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Describe this image"}
    ],
    "max_tokens": 256
  }'

When to Choose Rox Vision Max

• Visual understanding tasks
• Image analysis
• Multimodal applications

---