How do I choose the right AI model in Jylo?

For most tasks, start with GPT-4.1 - it's reliable and handles large documents well. Use GPT-5 when you need the deepest reasoning and enhanced citations.

What Are AI Models and Why Do They Matter?

AI models are the "brains" behind Jylo's analysis capabilities. Think of them like different types of assistants - some work quickly for everyday tasks, whilst others take more time to think deeply about complex problems.

Choosing the right model helps you get the best results for your specific documents. Don't worry - you can always switch between models to find what works best!

Our Top Models

GPT-4.1 Best for Everyday Work

Perfect for: Most document analysis tasks and Flow workflows

What it's great at:

• General purpose tasks
• Reviewing massive documents (up to about 750k words)
• Analysing many documents together with Assistant
• Reliable, consistent performance
• Structured Playbook workflows requiring predictable outputs

Tip: use this as your default model, especially for List, Yes/No Range, and Date prompts in Playbooks where consistency is critical

GPT-5 Best for Deep Reasoning and Research

Perfect for: Complex analysis requiring the deepest reasoning, detailed thinking and comprehensive citations

What it's great at:

• Superior reasoning that surpasses all previous models
• Superior citation generation - produces up to four times as many citations as GPT-4.1
• Extended output capacity (up to 128k tokens)
• Assistant conversations and text/generative prompt types
• Adaptive reasoning that switches between fast responses and deeper thinking

Note: GPT-5 uses a routing architecture that automatically adjusts its approach based on task complexity. Whilst powerful, this can produce unpredictable results in structured Playbook workflows. We recommend continuing to use GPT-4.1 for List, Yes/No, Range and Date prompts where consistency is essential.

Context capacity: 400k tokens (smaller than GPT-4.1 but with double the output capacity)

Additional Models

We offer several additional models for specific needs or budget considerations. These include options for multilingual content, mathematical problems, and various efficiency requirements, these include:

• o1
• o3
• o3-mini
• o4-mini
• gpt-4o
• Llama 3.3
• Mistral Large 2411
• Phi 3 Medium
• Phi 4
• Deepseek R1

For detailed comparisons, visit artificialanalysis.ai.

Need more help choosing? Contact us via support@jylo.ai.