Read the screenplay: FANNIEGATE — $7 trillion. 17 years. The biggest fraud in American capital markets.

Complete Guide · Updated March 2026

What is
Claude AI?

The complete guide to Anthropic's AI assistant — from a developer who built 800+ pages on this site using Claude Code. Models, pricing, capabilities, and honest takes on what it does well and where it falls short.

By Glen Bradford · Purdue Engineer, Salesforce Developer, Claude Power User

200K

tokens

Context Window

30+

supported

Languages

2021

Anthropic

Founded

800+

pages built

This Site

What is Claude?

Claude is an AI assistant built by Anthropic, an AI safety company founded in 2021 by Dario Amodei (CEO) and Daniela Amodei (President). Both are former executives at OpenAI who left to build an AI company with safety as a first principle, not an afterthought.

Unlike ChatGPT, which optimizes for engagement and breadth, Claude is designed around three core values: helpful, harmless, and honest. This isn't marketing. It's baked into the training process through Anthropic's Constitutional AI approach, where the model is trained to evaluate its own outputs against a set of principles.

In practice, this means Claude is less likely to make things up, more willing to say "I don't know," and more precise at following complex, multi-step instructions. It's the AI I use every single day to build software, write content, and analyze data.

The Claude Model Family

Anthropic offers three tiers of Claude, each optimized for different use cases. Think of it like choosing between a sports car, a sedan, and a motorcycle — they all get you there, but the ride is different.

Claude Opus

Most Capable

The heavy lifter for complex reasoning

Anthropic's most powerful model. Opus excels at complex analysis, nuanced writing, advanced mathematics, and multi-step reasoning tasks that stump other models. It's the model you reach for when accuracy matters more than speed.

Best For

  • Complex code architecture
  • Long-document analysis
  • Research synthesis
  • Advanced math & logic
  • Nuanced creative writing

Context Window

200K tokens

Speed

Slower, but worth the wait

Claude Sonnet

Best Balance

The sweet spot of speed and intelligence

The workhorse of the Claude family. Sonnet delivers near-Opus quality at significantly faster speeds and lower cost. Most developers and businesses run on Sonnet for day-to-day tasks. It handles coding, writing, and analysis with impressive capability.

Best For

  • Daily coding tasks
  • Content creation
  • Data analysis
  • Customer support
  • General-purpose assistant

Context Window

200K tokens

Speed

Fast and responsive

Claude Haiku

Fastest & Cheapest

Lightning-fast for high-volume tasks

The speed demon. Haiku is designed for tasks where latency and cost matter most. It's remarkably capable for its speed, handling summarization, classification, and simple Q&A at a fraction of the cost of larger models.

Best For

  • Real-time chat applications
  • Content classification
  • Quick summaries
  • High-volume API calls
  • Simple Q&A bots

Context Window

200K tokens

Speed

Near-instant responses

What Can Claude Do?

Claude isn't just a chatbot. It's a reasoning engine that handles a wide range of tasks with surprising depth.

Conversation & Analysis

Natural dialogue that actually understands context. Claude tracks long conversations, remembers details from earlier in the thread, and provides thoughtful, nuanced responses rather than generic AI slop.

Code Generation & Review

Writes production-quality code in dozens of languages. Claude understands codebases holistically, suggests architectural improvements, catches bugs, writes tests, and explains complex code in plain English.

Long Document Processing

With a 200K token context window, Claude can ingest entire books, legal contracts, or codebases in a single prompt. It finds patterns, contradictions, and key insights across hundreds of pages.

Writing & Creative Work

From blog posts to screenplays, Claude writes with genuine voice and style. It adapts tone, maintains consistency across long pieces, and produces content that doesn't read like it was written by a committee of robots.

Mathematics & Reasoning

Claude handles everything from algebra to advanced calculus, statistics, and formal logic. It shows its work step-by-step and can explain mathematical concepts at any level of complexity.

Vision & Image Understanding

Upload images and Claude can describe, analyze, and reason about them. It reads charts, interprets diagrams, extracts text from photos, and understands visual context with remarkable accuracy.

How to Use Claude

claude.ai (Web & Mobile)

Try Claude Free

The simplest way to start. Go to claude.ai, create a free account, and start chatting. The free tier gives you access to Claude Sonnet with a daily message limit. Upload files, paste documents, or just have a conversation. Available on web and mobile apps.

Claude Pro ($20/month)

Upgrade to Pro

Unlocks the full model lineup (Opus, Sonnet, Haiku), 5x more messages, Projects for organizing conversations, extended thinking mode for complex reasoning, and priority access during high-traffic periods. Worth every penny if you use Claude daily.

For developers building applications. Pay-per-token pricing, batch processing, and enterprise features. The API gives you raw access to all Claude models with fine-grained control over parameters like temperature, max tokens, and system prompts.

Claude Code (CLI for Developers)

Deep Dive Below

Anthropic's official command-line tool. This is how I built this website. Claude Code operates directly in your terminal, reads your files, runs commands, edits code, and manages Git workflows. It understands entire project structures, not just individual files.

Claude Code: A Deep Dive

How I built 800+ pages without writing most of the code myself

Claude Code is the tool that changed my workflow permanently. It's Anthropic's official CLI (command-line interface) that lets you use Claude directly in your terminal. Not as a chatbot that generates code snippets you copy-paste. As an agent that reads your files, understands your project structure, runs commands, and makes coordinated edits across multiple files.

This website — glenbradford.com — has over 800 pages. Blog posts, interactive tools, games, calculators, ranking pages, shrine pages. I built the vast majority of them by describing what I wanted to Claude Code in plain English. It writes the TypeScript, the Tailwind CSS, the JSON-LD schema, the metadata. It creates the file, adds it to the sitemap, and stages the git commit.

What Claude Code can do

  • Read & edit files across your project
  • Run terminal commands (build, test, lint)
  • Search codebases with regex & glob patterns
  • Create new files and directories
  • Stage commits and manage Git workflows
  • Understand project architecture holistically
  • Make coordinated changes across 50+ files
  • Debug errors by reading stack traces
  • Install dependencies and configure tools
  • Refactor code while preserving behavior

How is this different from GitHub Copilot? Copilot autocompletes the line you're writing. Claude Code understands your entire project and makes decisions at the architectural level. Copilot suggests; Claude Code executes. You can tell Claude Code "add a new page with this content, match the style of the other pages, add it to the sitemap and search index" and it will do all of that in one pass.

The page you're reading right now was built with Claude Code. I described what I wanted, and it wrote every line.

Anthropic's Approach: Constitutional AI

Most AI companies train their models with RLHF (Reinforcement Learning from Human Feedback) — humans rate outputs, and the model learns to produce higher-rated responses. Anthropic takes this further with Constitutional AI (CAI).

The idea is elegant: give the AI a set of principles (a "constitution") and have it critique its own outputs against those principles. Instead of just learning "humans liked this response," the model learns why a response is good or bad. This makes Claude more predictable, more principled in its refusals, and less susceptible to jailbreaks.

In practice, this means Claude will explain why it can't help with something rather than just refusing. It produces fewer hallucinations because it's trained to be honest about uncertainty. And it follows complex instructions more precisely because it understands the intent behind them, not just the surface pattern.

Anthropic also publishes research papers about their safety work, contributes to AI governance discussions, and has taken positions on responsible scaling. Whether you care about AI safety or not, the practical result is a model that's more reliable and less likely to go off the rails.

Claude Pricing

Claude offers a free tier, a Pro subscription, and usage-based API pricing. Here's how they compare.

Free

$0/forever
  • Access to Claude Sonnet
  • Limited daily messages
  • Basic file uploads
  • Standard response speed
  • Great for trying Claude out
Most Popular

Pro

$20/per month
  • Access to Opus, Sonnet, and Haiku
  • 5x more usage than free
  • Priority access during peak
  • Create and use Projects
  • Extended thinking for complex tasks
  • Early access to new features

API

Pay per token/usage-based
  • Haiku: ~$0.25 / 1M input tokens
  • Sonnet: ~$3 / 1M input tokens
  • Opus: ~$15 / 1M input tokens
  • Build custom applications
  • Batch processing discounts
  • Enterprise support available

Claude vs ChatGPT vs Gemini

An honest comparison. Every model has strengths. Here's how Claude stacks up against the competition in 2026.

FeatureClaudeChatGPTGemini
Context Window200K tokens128K tokens1M+ tokens
Coding AbilityExcellentVery GoodGood
Long DocumentsBest in classGoodExcellent
Instruction FollowingBest in classVery GoodGood
Creative WritingExcellentVery GoodGood
Safety & RefusalsPrincipled, rare over-refusalSometimes over-cautiousVariable
Internet BrowsingNot in base modelBuilt-inBuilt-in
Image GenerationNot availableDALL-E built-inImagen built-in
Developer CLIClaude Code (excellent)None officialNone official
Free TierYes (Sonnet)Yes (GPT-4o mini)Yes (Gemini Flash)

This comparison reflects the state of these models as of March 2026. AI moves fast — capabilities change with every model release. Your mileage may vary based on specific use cases.

What Claude is Best At

Long Document Analysis

Drop a 200-page contract into Claude and ask it to find every clause related to indemnification. It handles this better than any other model I've tested.

Production-Quality Code

Claude writes TypeScript, Python, Rust, Go, and more at a level that passes code review. It understands design patterns, not just syntax.

Nuanced Writing

Claude doesn't write like a robot. It adapts to tone, maintains voice consistency, and produces content that actually sounds human.

Complex Instruction Following

Give Claude a 15-point prompt with nested requirements and constraints. It handles multi-step instructions better than any competitor.

Honest Uncertainty

Claude will tell you when it's not sure about something instead of confidently making things up. This alone is worth the switch from other models.

Safety Without Annoyance

Claude rarely over-refuses. It handles mature topics thoughtfully rather than slapping a generic 'I can't help with that' on everything.

Where Claude Falls Short

No AI is perfect. Here's what Claude struggles with, so you know what you're getting into.

No Internet Browsing

Base Claude can't search the web or access real-time information. Its knowledge has a training cutoff date. You can work around this by pasting content into the conversation, but it's not the same as having built-in search like ChatGPT.

No Image Generation

Claude can analyze and describe images, but it can't create them. If you need AI-generated art or photos, you'll need a separate tool like Midjourney, DALL-E, or Stable Diffusion.

Hallucinations Still Happen

Claude hallucinates less than competitors in most benchmarks, but it still makes things up occasionally, especially for niche topics, specific dates, or obscure facts. Always verify critical information independently.

No Persistent Memory

Each conversation starts fresh. Claude doesn't remember what you talked about last week unless you use Projects to provide persistent context. This is a deliberate privacy choice, but it can be inconvenient.

Rate Limits on Free Tier

The free tier has a relatively low daily message limit. Heavy users will hit the cap quickly, especially during peak hours. Pro ($20/month) significantly increases this, but it's still not unlimited.

Get Glen's Musings

Occasional thoughts on AI, Claude, investing, and building things. Free. No spam.

Unsubscribe anytime. I respect your inbox more than Congress respects property rights.

Glen's Experience with Claude

I've been using Claude daily since early 2025. I started with the web interface, moved to Pro, and then discovered Claude Code. That last one changed everything.

This website — glenbradford.com — has over 800 pages. Blog posts, interactive games (a 3D kitesurfing simulator), financial calculators, ranking pages, comedy pages, screenplays, shrine pages for people in my life. The vast majority were built with Claude Code.

My workflow: I describe what I want in natural language. Claude Code reads the existing codebase, understands the patterns (Next.js 16, Tailwind 4, TypeScript, my component library), and writes the complete implementation. I review, iterate, and ship. What used to take me a full day takes under an hour.

Before Claude, I used GitHub Copilot. Copilot is good at autocompleting the line you're writing. Claude Code is good at building the entire feature. The difference is like comparing spell-check to having a coauthor.

I'm a Purdue-trained engineer who ran a hedge fund, worked as a Salesforce developer, and now builds software with AI. I've tried every coding assistant on the market. Claude is the one I keep coming back to. This page you're reading was built with Claude.

The Future of Claude

Anthropic is investing heavily in several directions that will shape Claude's evolution:

Computer Use

Claude can already control a computer by moving the mouse, clicking buttons, and typing text. This capability is evolving from demo to production-ready, enabling Claude to use any software a human can use.

Agent Capabilities

Multi-step autonomous workflows where Claude breaks down complex tasks, executes them independently, handles errors, and delivers results. Think: 'research these 50 companies and create a comparison spreadsheet.'

Longer Context Windows

200K tokens is already industry-leading, but Anthropic is pushing toward even larger context windows. Imagine analyzing an entire codebase or a year's worth of emails in a single conversation.

Improved Reasoning

Extended thinking and chain-of-thought reasoning are getting deeper and more reliable. Future Claude models will tackle problems that require sustained multi-step reasoning across minutes or hours.

Better Tool Integration

Claude is gaining the ability to use external tools, APIs, and services natively. This means real-time data access, code execution, and integration with enterprise systems will become seamless.

Safety at Scale

Anthropic's responsible scaling policy ensures that as Claude gets more powerful, safety measures scale proportionally. This is the key differentiator from labs that ship first and safety-check later.

Frequently Asked Questions

What is Claude AI?

Claude is an AI assistant created by Anthropic, a company founded by former OpenAI researchers Dario and Daniela Amodei. Claude is designed to be helpful, harmless, and honest. It can hold conversations, write code, analyze documents, create content, solve math problems, and more. It's available as a free chatbot at claude.ai, as a paid Pro subscription, and as an API for developers.

Is Claude AI free to use?

Yes, Claude has a free tier at claude.ai that gives you access to Claude Sonnet with a limited number of daily messages. For heavier usage, Claude Pro costs $20/month and includes access to all models (Opus, Sonnet, Haiku), 5x more usage, and features like Projects and extended thinking.

How is Claude different from ChatGPT?

Claude tends to excel at following complex instructions, processing long documents (200K token context window), coding tasks, and producing nuanced writing. ChatGPT has advantages in internet browsing, image generation via DALL-E, and a larger plugin ecosystem. Claude is generally considered more precise with instructions and less likely to hallucinate, while ChatGPT has broader integrations.

What is Claude Code?

Claude Code is Anthropic's official CLI (command-line interface) tool for developers. It lets you use Claude directly in your terminal to write code, edit files, run commands, search codebases, and manage Git workflows. It understands entire project structures and can make coordinated changes across multiple files. Glen Bradford built over 800 pages of glenbradford.com using Claude Code.

What is Anthropic's Constitutional AI?

Constitutional AI (CAI) is Anthropic's approach to AI safety. Instead of relying solely on human feedback to train models, CAI gives the AI a set of principles (a 'constitution') and has the model critique and revise its own outputs. This produces an AI that's helpful while being less likely to produce harmful content, with transparent reasoning about why it refuses certain requests.

What is Claude's context window?

Claude supports a 200K token context window, which is roughly equivalent to 150,000 words or about 500 pages of text. This means you can paste entire books, legal contracts, codebases, or research papers into a single conversation and Claude can analyze them holistically. Some Claude models support even larger context windows for specialized use cases.

Can Claude browse the internet?

Base Claude does not have internet browsing capability. Its knowledge comes from its training data, with a cutoff date. However, you can paste text, URLs' content, or documents into Claude for analysis. Some integrations and tools (like Claude Code) give Claude the ability to fetch web content when needed.

Is Claude good for coding?

Claude is widely regarded as one of the best AI models for coding tasks. It excels at understanding large codebases, writing production-quality code, debugging, writing tests, and explaining complex code. Claude Code, the CLI tool, takes this further by letting developers use Claude directly in their terminal with full file system access. Multiple benchmarks rank Claude at or near the top for coding capability.

Who founded Anthropic?

Anthropic was founded in 2021 by Dario Amodei (CEO) and Daniela Amodei (President), along with several other former OpenAI researchers. They left OpenAI to build an AI company with safety at its core. Anthropic has raised billions in funding and is one of the leading AI labs, alongside OpenAI and Google DeepMind.

What can't Claude do?

Claude cannot browse the internet in real-time, generate images, access external tools or APIs on its own, remember previous conversations (each chat starts fresh), or execute code in a sandbox environment through the web interface. It can also hallucinate facts, especially about niche topics or recent events after its training cutoff. It's a text-based reasoning engine, not an omniscient oracle.

Keep Exploring