Smart Auto-Routing

Intelligent Model Selection

SOHAM automatically analyzes your queries and routes them to the most appropriate AI model for optimal results.

How Smart Routing Works

Advanced query analysis for automatic model selection

1

Query Analysis

AI analyzes your message content, context, and intent

2

Model Matching

Selects the best-suited model from 13+ available options

3

Optimized Response

Delivers the highest quality answer for your specific query

🎯 Key Benefits

  • • No manual model selection required
  • • Always get the best model for each task
  • • Seamless switching between specialized models
  • • Optimal performance without complexity

Routing Categories

Smart routing identifies these types of queries and routes them to specialized models:

Mathematical Queries

Equations, calculations, and mathematical problem solving

Example Queries:

  • "Solve x² + 5x + 6 = 0"
  • "What is the derivative of sin(x)?"
  • "Calculate the area of a circle with radius 5"
  • "Explain the quadratic formula"

Target Models:

WizardMath 70B
Qwen 2.5 (Math capable)
Llama models with math training

Programming & Code

Software development, debugging, and coding assistance

Example Queries:

  • "Write a Python function to sort a list"
  • "Debug this JavaScript error"
  • "Explain React hooks"
  • "Best practices for API design"

Target Models:

DeepSeek V3.2
Coding-specialized models
Programming-trained variants

General Conversation

Everyday questions, explanations, and casual chat

Example Queries:

  • "Explain how photosynthesis works"
  • "What are the benefits of exercise?"
  • "Tell me about the history of Rome"
  • "How do I improve my writing skills?"

Target Models:

Llama 3.1 models
General purpose models
Conversational variants

Visual & Multimodal

Image analysis, visual questions, and multimodal tasks

Example Queries:

  • "Describe this image"
  • "What do you see in this picture?"
  • "Analyze this chart or graph"
  • "Read text from this image"

Target Models:

Gemini 2.5 Flash
Gemini Flash Latest
Vision-capable models

Query Analysis Process

What the AI Looks For

Key indicators that determine model routing decisions

Content Analysis

Mathematical Keywords

solve, calculate, equation, derivative, integral

Routes to math models
Programming Terms

function, variable, debug, code, algorithm

Routes to coding models
Visual References

image, picture, photo, visual, chart

Routes to multimodal models

Context Analysis

Question Structure

Step-by-step requests, Explanation needs, Problem format

Influences model choice
Conversation History

Previous topic context, User expertise level, Ongoing discussion

Maintains consistency
Complexity Level

Simple vs complex queries, Technical depth needed, Detail requirements

Selects appropriate model

Routing Examples

Query:

"Solve the equation 2x + 5 = 15"

Mathematics
ANALYSIS

Mathematical equation with 'solve' keyword

SELECTED MODEL

Math-specialized model (e.g., WizardMath)

REASONING

Query contains mathematical equation and explicit solve request

Query:

"Write a Python function to reverse a string"

Programming
ANALYSIS

Programming request with specific language

SELECTED MODEL

Coding model (e.g., DeepSeek V3.2)

REASONING

Contains programming language and function creation request

Query:

"What's the weather like today?"

General + Search
ANALYSIS

General information request

SELECTED MODEL

General model with search capability

REASONING

Requires real-time information, routes to search-enabled model

Query:

"Explain the concept of machine learning"

Education
ANALYSIS

Educational explanation request

SELECTED MODEL

Conversational model (e.g., Llama 3.1)

REASONING

General knowledge explanation, benefits from conversational model

Manual Model Selection

When and how to override smart routing

✅ When to Use Auto Mode

  • • Most general usage scenarios
  • • Mixed conversation topics
  • • When you want optimal results
  • • Learning and exploration
  • • First-time users

⚙️ When to Use Manual Selection

  • • Consistent model behavior needed
  • • Testing specific model capabilities
  • • Advanced users with preferences
  • • Specialized workflows
  • • Research or comparison purposes

💡 Recommendation

Start with Auto mode to experience the full power of smart routing. You can always switch to manual selection later if you have specific model preferences or requirements.

Performance Benefits

How smart routing improves your experience

Higher Accuracy

Specialized models perform better on their target tasks

Up to 40% better results

Faster Responses

Right-sized models for each query type

Optimized speed

Better Context

Models trained for specific domains understand context better

Improved relevance

Seamless Experience

No manual switching or configuration needed

Zero complexity

Advanced Routing Features

Sophisticated capabilities of the smart routing system

Context Preservation

Maintains conversation context when switching between models

Example: Math question followed by explanation request uses consistent context

Fallback Handling

Automatically tries alternative models if the first choice fails

Example: If specialized model is unavailable, falls back to general model

Multi-Modal Detection

Identifies when queries involve multiple types of content

Example: Code with math components routes to models capable of both

Learning Adaptation

Routing improves based on successful interactions

Example: System learns which models work best for your query patterns

Previous

Web Search

Next

Jarvis Mode