Assistant List
This document explains the assistant features of AI-Starter. Assistants help users create content tailored to their needs by combining various generative AI models and options.
Supported Models
AI-Starter supports a variety of generative AI models, with available models varying by environment.
Model Characteristics
The following is based on evaluations within the development team, with each model having different strengths and characteristics.
- Nova series: Models available at a low price.
- Claude series: Models with high text comprehension and stable performance.
- Gemini series: Models with excellent cost-performance and information processing capabilities.
- OpenAI series: Highly versatile models incorporating cutting-edge technology.
List of Supported Models
Here's the English translation:
- AWS
- Claude 4 Sonnet
- Claude 4 Thinking
- Claude 4 Opus
- Claude 3.5 Haiku
- Claude 3 Haiku
- Nova Micro
- Nova Lite
- Nova Pro
- Nova Premier
- Azure
- Azure OpenAI
- GPT-4o mini
- GPT-4o
- GPT-4.1
- o4 mini
- o1
- DALL·E 3
- Azure AI Foundry
- Grok 3 (Trial Offering)
- Azure OpenAI
- Google
- Gemini 2.0 Flash
- Gemini 2.0 Flash-Lite
- Gemini 2.5 Flash Preview
- Gemini 2.5 Pro Preview
- Gemini 2.0 Flash Image Generation
- Gemini 2.0 Flash Experimental
- Imagen 4
- Imagen 4 Ultra
- Imagen 3
- OpenAI
- GPT-4o mini
- GPT-4o
- GPT-4.1 nano
- GPT-4.1 mini
- GPT-4.1
- GPT-4o Search mini Preview
- GPT-4o Search Preview
- ChatGPT-4o latest (Experimental Model)
- o4 mini
- o3
- DALL·E 3
- GPT Image 1
Notes
- Some models support video input, but this is currently not supported in AI-Starter.
- Preview and experimental models may be discontinued without notice at the provider's discretion. Please be aware of this in advance.
Related Information
- AWS
- Azure
- OpenAI
Options
Assistants can use various options in combination with generative AI models. These options enable more advanced content creation. For enabling settings, please contact the support desk through your internal administrator.
| Option Name | Description |
|---|---|
| KendraRetrieval | A RAG (Retrieval Augmented Generation) option using Amazon Kendra. It can be flexibly used with any model. |
| BedrockKnowledgeBasesRetrieval | An option that allows flexible RAG construction using Amazon Bedrock Knowledge Bases. |
| BedrockKnowledgeBasesRetrieveAndGenerate | A simple RAG option using Amazon Bedrock Knowledge Bases for consistent search and generation. |
| OpenAIResponsesAPI | An option to build agents utilizing OpenAI's built-in tools. |
| GoogleDriveFiles (Beta) | An option that allows incorporating content from specified Google Drive files into conversations. |
| WebPages | A scraping option for retrieving information from web pages. |
| StorageFiles | An option to use locally uploaded files as input. |
KendraRetrieval
The KendraRetrieval option allows you to perform RAG (Retrieval Augmented Generation) using Amazon Kendra.
With RAG, when a user asks a question, documents containing information needed for the answer are searched, and those documents are added to the input (prompt) for the generative AI to generate an answer based on the information the user is seeking. This enables answers based on specific internal information such as internal system manuals or FAQs.
To use this, you need to create an Amazon Kendra index and specify the index ID to the assistant.
BedrockKnowledgeBasesRetrieval
The BedrockKnowledgeBasesRetrieval option allows you to perform RAG (Retrieval Augmented Generation) using Amazon Bedrock Knowledge Bases. You can use Knowledge Bases for searching your custom data and any model for generating responses. This enables flexible construction of high-precision RAG.
To use this, you need to create Knowledge Bases and specify the knowledge base ID to the assistant.
BedrockKnowledgeBasesRetriveAndGenerate
The BedrockKnowledgeBasesRetriveAndGenerate option allows you to perform RAG (Retrieval Augmented Generation) using Amazon Bedrock Knowledge Bases. You can easily implement RAG by using Knowledge Bases for everything from search to response.
To use this, you need to create Knowledge Bases and specify the knowledge base ID and a model in the same region to the assistant.
OpenAIResponsesAPI
The OpenAIResponsesAPI option allows you to build agents utilizing OpenAI's built-in tools. Currently, AI-Starter supports two tools: web search and file search.
To use this, you need to select built-in tools and specify a model that supports the Responses API.
GoogleDriveFiles (Beta)
The GoogleDriveFiles option allows you to use Google Drive files as input for the assistant. Currently, the following files can be imported:
- Google Documents
- First sheet of Google Spreadsheets
To use this, you need to specify the file ID from Google Drive.
GoogleCustomSearchRetrieval
The GoogleCustomSearchRetrieval option allows you to get answers based on Google custom search results. To use it, you need to set the following information in environment variables:
- Project ID linked to Vertex AI
- Vertex AI usage region
- API key for the service account linked to Vertex AI
- Custom search engine ID
Details
Notes on using custom search
Google Custom Search collects data during requests and responses. When using it, it is recommended to advise users not to input sensitive information. Custom Search Engine Terms of ServiceGoogle and its wholly-owned subsidiaries may retain and use information collected through your use of the Service in accordance with the terms of the Privacy Policy (http://www.google.com/privacy.html or such other URL as Google may provide from time to time).
WebPages
The WebPages option allows information retrieval through web page scraping. This is a technique that analyzes the HTML structure of web pages and extracts necessary information.
With the WebPages option, you can retrieve information in the following ways:
- Collect information from pre-specified web page URLs
- Collect information from URLs entered by users
StorageFiles
The StorageFiles option allows you to use files uploaded locally as input for the assistant. For example, you can upload a PDF file and ask questions about its contents. You can choose which settings to enable for each assistant.
For details, refer to Attachment File Support Status.
Recommended Models
| Content | Recommended Model |
|---|---|
| General Q&A | Claude 4 Sonnet |
| Advanced Q&A | Claude 4 Opus |
| Creative Writing | Claude 4 Sonnet |
| Coding | Claude 4 Sonnet |
| Advanced Coding | Claude 4 Opus |
| Business Document Creation Support | Claude 4 Sonnet |
| Translation Support | Gemini 2.0 Flash |
| Long Text Comprehension | Gemini 2.0 Flash |
| Image-related Questions | Gemini 2.0 Flash |
| Text File-related Questions | Gemini 2.0 Flash |
| PDF File-related Questions | Gemini 2.0 Flash |
| Spreadsheet File-related Questions | Claude 4 Sonnet |
| Just Want to Try Something | Nova Premier |
Additional Information
Generative AI Model Comparison
The following external sites are useful for comparing generative AI models. However, as information is updated regularly, always check for the latest information and reliability.