Skip to main content

Choosing the right AI model

Select the optimal AI model for your specific tasks and use cases

Jake Rosenthal avatar
Written by Jake Rosenthal
Updated over 2 weeks ago

Cassidy gives you access to all the leading AI models and automatically recommends the best options for different scenarios. The platform's intelligent model selection system helps you choose the right tool for each task while keeping recommendations updated as new models are released.

Using models across Cassidy features

In Assistants

Choose any foundational model as the base for your custom Assistants. You can also access default Assistants that use foundational models without being trained on your organization's Knowledge Base.

In Chat

The Chat interface allows you to easily switch between different AI models and Assistants during conversations, giving you flexibility to use the best model for each part of your conversation.

In Workflows

Workflows can leverage different models through various actions, including the "Generate Text" integration which allows you to select the AI model that best fits each automation step.

Staying current with recommendations

Cassidy's "Recommended Models" section is continuously updated as new models are released and evaluated. The platform automatically surfaces the best-performing, most cost-effective models for your needs, and model cost reductions from providers are automatically passed on to users.

The pace of AI advancement is accelerating, with new models and improvements arriving regularly. Cassidy ensures you always have access to the latest capabilities without needing to track releases manually.

Understanding model types and their strengths

GPT (OpenAI)

Best for general assistant applications, everyday tasks, and Q&A. GPT models are reliable for retrieving information and handling a wide range of business queries. They excel at automation and general knowledge tasks.

Claude (Anthropic)

Preferred for creative writing, sales, and marketing content that needs a more human-like or nuanced touch. Claude is ideal for drafting emails, marketing copy, and communication-heavy tasks that require natural-sounding language.

Gemini (Google)

Ideal for use cases that require processing large amounts of information or synthesizing data from multiple sources, thanks to its larger context window. Perfect for deep research, complex data analysis, and summarizing lengthy documents.

Model categories by capability

Mini Models

  • Purpose: Designed for speed and cost-efficiency

  • Best for: High-volume, lower-value, or latency-sensitive tasks like customer support, triage, and operational reviews

  • Benefits: Fast responses, cost-effective for routine automation

  • Note: Latest mini models often match or exceed the performance of previous-generation full models

High-Reasoning Models

  • Purpose: Engineered for complex, multi-step reasoning and high-stakes decision-making

  • Best for: Legal analysis, financial workflows, STEM problem-solving, and scenarios where accuracy and auditability are critical

  • Characteristics: "Think" longer before responding, break down problems into stepwise solutions, and self-check their work

  • Trade-offs: Slower and more expensive, but essential for demanding analytical tasks

General-Purpose Models

  • Purpose: Balanced performance for everyday use

  • Best for: Q&A, creative writing, and moderately complex tasks

  • Benefits: Faster and more cost-effective than reasoning models for simple tasks

  • Usage: Can be prompted to "think step by step" but don't default to deep reasoning

Model selection best practices

Match the model to the task

  • Use mini models for summarization, formatting, and routine workflow steps

  • Use high-reasoning models for complex analysis, planning, or when accuracy is paramount

  • Use general-purpose models for everyday tasks that don't require deep reasoning

Mix and match in Workflows

For optimal efficiency, combine different models within the same Workflow:

  • Use mini models for straightforward steps like data formatting

  • Reserve high-power models for generating final content or complex analysis

  • Apply reasoning models for multi-step analytical work or critical decision points

Consider your specific use case

  • GPT: Quick answers, document retrieval, general automation

  • Claude: Natural-sounding writing, communication tasks, creative content

  • Gemini: Large-scale data processing, research synthesis, lengthy document analysis

By leveraging Cassidy's comprehensive model selection and intelligent recommendations, you can ensure optimal performance and cost-efficiency across all your AI-powered tasks, whether you're handling routine automation or complex analytical work.

Did this answer your question?