---
title: "Selecting an AI Model"
canonical: "https://wiki.yellowfinbi.com/space/yfcurrent/2228893/Selecting%20an%20AI%20Model"
format: markdown
---
Yellowfin supports connections to a range of cloud-hosted AI models. The aim of this is to provide a range of choices to Customers that best align with their specific needs. The currently supported models are:

### OpenAI

- GPT-4o mini
- GPT-4o
- GPT-4.1
- o3-mini
- GPT-5.2
- GPT-5.4
- GPT-5.5
- GPT-5.6 Sol
- GPT-5.6 Terra
- GPT-5.6 Luna

### Azure OpenAI

- GPT-4o
- GPT-4.1
- O3-mini
- GPT-5.2
- GPT-5.4 mini
- GPT-5.4
- GPT-5.5
- GPT-5.6 Sol
- GPT-5.6 Terra
- GPT-5.6 Luna

### Google Gemini

- Gemini 2.5 Flash
- Gemini 2.5 Pro
- Gemini 3 Flash Preview
- Gemini 3.1 Pro Preview
- 3.5 Flash
- 3.6 Flash

### Anthropic Claude

- Sonnet 4.5
- Clause Opus 4.5
- Claude Haiku 4.5
- Clause Opus 4.6
- Claude Sonnet 4.6
- Claude Opus 4.7
- Fable 5
- Opus 4.8
- Opus 5
- Sonnet 5

Yellowfin has conducted testing against the supported models. Each model provides a range of different outcomes, including different performance, accuracy and cost. **We strongly recommend that Customers undertake their own testing against different models, and select the model most appropriate for their needs.**

Our analysis has shown excellent performance with the latest versions of ChatGPT, Claude and Gemini Pro. These models offer the best outcome in terms of quality of responses, speed and cost. We recommend updating as newer models become available to maintain optimal results.

Other important things to note:

- It is possible to mix and match models based on feature. For example, OpenAI o3 can be used for AI-enabled Assisted Insights (due to the richness of narratives returned) but OpenAI GPT-4o used for NLQ, due to its speed and accuracy. You can even mix models from different model providers.
- Responses from models can be inconsistent. Large Language Models (LLMs) are predictive models by their nature, and unlike deterministic code that is always guaranteed to return the same output given the same input. AI-models are designed to provide assistance to make certain tasks like data analysis, easier.
- Different connections can be created at Client Org level. This enables Customers who run multi-tenant deployments, the ability to have dedicated model connections per client, and thus be easier able to account for costs across the client base.