AI Services in Generative AI (LLM)

Large language models capable of generating natural language, images, code, and moreFeaturing 16 AI services in total.

マナス

Manus

A fully autonomous AI agent that handles end-to-end tasks spanning research, report writing, data organization, and website generation. Simply provide a goal and the AI independently plans, executes, and validates its way to completion. Its multi-agent architecture pairs a Planner, an Executor, and a Validator working in concert.

# AIエージェント# 生成AI（LLM）# 自動化・省力化# 業務効率化

グロック

Grok

A conversational AI assistant developed by Elon Musk's xAI. Its key strengths are access to real-time information from X (Twitter) and a response style that blends accurate, up-to-date answers with humor. Originally exclusive to X premium subscribers, it is now also accessible standalone at grok.com. The latest model is Grok 3.

# 生成AI（LLM）# AIチャットボット# テキスト生成# 画像生成# AIエージェント# 日本語対応

パープレキシティ

Perplexity

An answer-first AI search engine that cites source URLs alongside every response. Supports multiple LLMs including GPT, Claude, Gemini, and Sonar, with file upload and image generation available on Pro plans. Offers Perplexity Pages for generating summary articles and an AI assistant experience — a rapidly growing service redefining how people search.

# 検索・レコメンド# 検索エンジン・RAG# AIエージェント# 要約・翻訳# 生成AI（LLM）# 日本語対応

ジェミニ

Gemini

A multimodal AI developed by Google, capable of processing a wide range of information including search queries, images, and documents.

# 生成AI（LLM）# AIチャットボット# 要約・翻訳# テキスト生成# AIエージェント# 日本語対応

チャット ジーピーティー

ChatGPT

A conversational AI developed by OpenAI. Using the latest models like GPT-4o and o3, it supports text, image, and voice conversations, file analysis, web search, and code generation and execution. 'Projects' enable custom instructions, while the Canvas feature enables collaborative editing of text and code. Widely integrated into external applications via API.

# 生成AI（LLM）# AIチャットボット# 要約・翻訳# テキスト生成# AIエージェント# 日本語対応

クロード

Claude

A conversational AI assistant developed by Anthropic. Handles a wide range of tasks including comprehending and generating long-form content, writing and reviewing code, and analyzing documents. Designed with safety (Constitutional AI) and honest responses as core principles. The latest Claude 4 model achieves top-tier benchmark performance in coding, mathematics, and reasoning. Available at claude.ai, via API, and through the Claude app.

# 生成AI（LLM）# AIチャットボット# 要約・翻訳# テキスト生成# AIエージェント# 日本語対応

ヴァンナエーアイ

Vanna AI

An open-source (MIT) RAG framework for Text-to-SQL that generates accurate SQL from natural language questions. Train it on your database DDL, documentation, and SQL examples, and it uses Agentic Retrieval to produce high-precision queries. Supports major databases including PostgreSQL, Snowflake, and BigQuery, as well as leading LLMs from OpenAI, Anthropic, and others.

# 検索エンジン・RAG# コード生成# AIエージェント# 生成AI（LLM）

アジェンタ

Agenta

An open-source (MIT) LLMOps platform that unifies prompt management, LLM evaluation, observability (tracing), and a prompt playground in a single tool. Supports Git-style versioning, branching, and environment switching for prompts, enabling developers and non-engineers to collaborate on LLM applications in production. Free for self-hosting; a cloud-hosted version is also available.

# AIモデル構築・管理# AIエージェント# API・SDK# 生成AI（LLM）

クウェン

Qwen

The large language model "Qwen" (Tongyi Qianwen) developed by Alibaba, along with its conversational AI interface Qwen Chat. Supports multimodal input—text, images, and video—and offers free access to chat, image generation, video generation, web search, document processing, and code generation. Many models are released as open source, with full support for Japanese input and output.

# 生成AI（LLM）# AIチャットボット# テキスト生成# 画像生成# コード生成# 日本語対応

ミストラル

Mistral

A family of open-source and proprietary large language models (LLMs) developed by French AI startup Mistral AI, along with an API platform for accessing them. Ranges from the fast and lightweight Mistral 7B and Mixtral (MoE architecture) to Codestral for coding and Mistral Reasoning for advanced reasoning tasks. Supports both self-hosting and commercial API access.

# 生成AI（LLM）# AIチャットボット# テキスト生成# 要約・翻訳# AIエージェント# 日本語対応

グーグル エーアイ スタジオ

Google AI Studio

A free, developer-facing web tool for building and testing prototypes with Google's Gemini models. Provides an intuitive environment for prompt design, testing, and iteration, with seamless integration into production apps via the Gemini API. Supports multimodal input including text, images, audio, and video.

# 生成AI（LLM）# AIモデル構築・管理# API・SDK# AIエージェント

ディープシーク

DeepSeek

A series of large language models (LLMs) developed by Chinese AI company DeepSeek. The open-source DeepSeek-R1 attracted wide attention for matching or surpassing GPT-4o and Claude 3.5 Sonnet on reasoning benchmarks. Known for an efficient Mixture of Experts (MoE) architecture and open weights available for commercial use.

# 検索エンジン・RAG# 生成AI（LLM）

ステーブルディフュージョン

Stable Diffusion

An open-source AI image generation model that produces high-quality images from text prompts. Developed by Stability AI, it can be run locally and supports advanced customization through tools like ComfyUI and Automatic1111. Model weights are publicly available, enabling free fine-tuning and custom model creation.

# 画像生成# 生成AI（LLM）# アイデア支援／創造ツール# 日本語対応

リプリット

Replit

A cloud IDE that runs in the browser. Supports 60+ programming languages, with code execution, sharing, and publishing entirely in the browser. Includes Replit AI for code generation, explanation, and debugging. Scales from programming beginners to production deployment. The Replit Agent can also build and deploy applications from plain-text instructions.

# コード生成# 生成AI（LLM）# AIエージェント

アドビファイアフライ

Adobe Firefly

Adobe's generative AI toolset that creates images, vectors, and design elements from text prompts alone. Integrates with Photoshop, Illustrator, and other Adobe apps, with support for commercial use.

# 画像生成# デザイン生成# 生成AI（LLM）# 日本語対応

フレームパック

FramePack

An AI video generation model (ImageFramePack) that generates video interpolating between a designated first frame and last frame. Achieves more accurate motion generation than traditional video outpainting or image morphing, and can generate long-form video (1 minute or more) even on limited VRAM. An open-source model developed by lllysviel (Lvmin Zhang).

# 動画生成# 生成AI（LLM）# 品質向上・クリエイティブ支援