Free downloadsdelivered by email. Pick a document and we email you the PDF.
Claude in practice Anthropic's assistant, known for careful reasoning, strong writing, and the best end-to-end coding experience of the major chatbots. A one-model brief: what it is best for, how to use it, pricing, and example prompts. ChatGPT in practice OpenAI's assistant and the most widely used chatbot, a generalist with the broadest feature set: voice, image generation, web browsing, data analysis, and custom GPTs. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Gemini in practice Google's AI assistant, natively multimodal, with the largest context window of the major models and deep integration across Search, Workspace, and Android. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Grok in practice xAI's assistant, built into X (formerly Twitter), fast, willing to be blunt, and uniquely tied to real-time social data. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Mistral in practice A European AI lab known for strong open-weight models you can self-host, plus Le Chat, a fast assistant with a real data-control story. A one-model brief: what it is best for, how to use it, pricing, and example prompts. DeepSeek in practice A Chinese lab whose open-weight models, especially the R1 reasoning model, matched frontier quality at a fraction of the cost, and reset expectations for AI pricing. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Qwen in practice Alibaba's open-weight model family, one of the broadest and most-downloaded, spanning tiny on-device models to large multimodal and coding variants. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Perplexity in practice An AI answer engine: ask a question, get a direct answer written from a live web search, with numbered citations to every source. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Llama in practice Meta's open-weight model family, the most widely downloaded foundation for self-hosted and on-device AI, free to run and fine-tune under a community license. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Microsoft Copilot in practice Microsoft's assistant, embedded across Windows, Microsoft 365, and Edge, built on OpenAI's models and grounded in your organization's own work data. A one-model brief: what it is best for, how to use it, pricing, and example prompts. GitHub Copilot in practice The AI pair programmer built into your editor, offering inline code completion, chat about your repo, and an agent mode across VS Code, JetBrains, and GitHub. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Kimi in practice Moonshot AI's assistant, known for very long context and, with the open-weight Kimi K2 model, strong agentic and coding ability at low cost. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Cohere Command in practice Cohere's enterprise model family, built for retrieval-augmented generation, tool use, and multilingual work, with private deployment inside your own cloud. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Ernie in practice Baidu's model family and assistant, one of China's most established, with strong Chinese-language and multimodal ability and an open-weight ERNIE 4.5 release. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Doubao in practice ByteDance's assistant and model family, one of China's most used, known for very low-cost API pricing through Volcano Engine and strong multimodal features. A one-model brief: what it is best for, how to use it, pricing, and example prompts. MiniMax in practice A Chinese lab whose models pair very long context with strong audio and video generation, with an open-weight text model you can self-host. A one-model brief: what it is best for, how to use it, pricing, and example prompts. GLM (Z.ai) in practice Zhipu AI's open-weight GLM family, now led by the agentic coding model GLM-5.2: a permissively licensed, self-hostable model built for long-horizon, repository-scale software engineering. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Amazon Nova in practice Amazon's own foundation models, served through AWS Bedrock and tuned for cost-efficiency, with text, image, and video variants for builders on AWS. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Jamba in practice AI21 Labs' open-weight model family, built on a hybrid Mamba-Transformer architecture for efficient, very long-context work. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Granite in practice IBM's family of open, Apache-licensed enterprise models, focused on efficient sizes, code, and governance for business and regulated use. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Falcon in practice The open-weight model family from the Technology Innovation Institute in Abu Dhabi, an early open-model leader with permissive licensing and self-hostable sizes. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Sakana Fugu in practice Sakana AI's multi-agent orchestration system, delivered as a single API that assembles and coordinates many models to run complex, multi-step tasks autonomously. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Devin in practice Cognition AI's autonomous software engineer: an agent that plans and ships code in its own cloud environment, with a shell, editor, and browser, and reports back like a teammate. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Manus in practice A general autonomous agent that runs multi-step tasks in the cloud on its own, from research and data work to building simple sites and documents, and hands you the finished result. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Genspark in practice A super agent that plans a task and uses a toolbox of its own, from web search and document generation to slides, sheets, and even automated phone calls, to deliver a finished result. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Salesforce Agentforce in practice Salesforce's platform for building and deploying autonomous agents grounded in your CRM data, with a reasoning engine, guardrails, and consumption-based pricing for the enterprise. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Google Jules in practice Google's asynchronous coding agent: powered by Gemini, it clones your repository into a cloud machine, works on a task on its own, and opens a pull request for review. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Amazon Nova Act in practice Amazon's model and SDK for building reliable agents that take actions in a web browser, aimed at breaking complex workflows into small, dependable steps developers can compose. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Devstral in practice Mistral's open-weight coding model built for agentic software engineering: tuned to run inside agent harnesses, edit across files, and resolve real GitHub issues, with weights you can self-host. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Cursor in practice An AI-native code editor from Anysphere: a familiar editor with deep AI built in, from inline completion to an agent mode that plans and edits across your whole project. A one-model brief: what it is best for, how to use it, pricing, and example prompts. Replit Agent in practice An agent that builds and deploys a working app from a plain-language prompt, inside Replit's cloud environment, so you can go from idea to a live URL without managing setup. A one-model brief: what it is best for, how to use it, pricing, and example prompts. OpenHands in practice An open-source autonomous software-development agent from All Hands AI: it runs in a sandboxed environment with a terminal, editor, and browser, and you can self-host it or use the hosted cloud. A one-model brief: what it is best for, how to use it, pricing, and example prompts. ChatGPT Agent in practice OpenAI's agentic mode inside ChatGPT: given a goal, it works in its own virtual computer, browsing the web and using tools to carry out multi-step tasks and return a finished result. A one-model brief: what it is best for, how to use it, pricing, and example prompts. The AI Agents Field Guide What an AI agent actually is, when to deploy one, the failure modes, and the guardrails that keep it safe. Choosing an AI Agent The 18 agentic products compared: surfaces, best-for, pricing, and a selection path by scenario. The AI Concepts Glossary Plain-language explanations of the AI concepts that come up in real projects.
