Claude 3
By Anthropic
Claude 3 is a family of large language models released by Anthropic in March 2024, comprising three tiers — Haiku, Sonnet, and Opus, in ascending order of capability and cost — that introduced native vision input and set new benchmark…
Definition
Claude 3 is a family of large language models released by Anthropic in March 2024, comprising three tiers — Haiku, Sonnet, and Opus, in ascending order of capability and cost — that introduced native vision input and set new benchmark performance for Anthropic's models at the time.
Overview
Claude 3 was Anthropic's third major model generation, released in March 2024 as a step change from Claude 2, both in benchmark performance and in product design. The three-tier structure — Haiku (fastest and cheapest), Sonnet (balanced), and Opus (most capable) — let developers and Claude.ai users pick the right cost/latency/capability tradeoff for a given task, a tiering approach Anthropic has continued in later generations. At launch, Claude 3 Opus was positioned as competitive with or exceeding GPT-4 on several published benchmarks, marking Anthropic's most direct challenge yet to OpenAI's flagship model. Claude 3 introduced native multimodal (vision) input across all three tiers, allowing the models to process images such as charts, photos, screenshots, and documents alongside text — a capability Claude 2 lacked. The models also expanded context window support up to 200K tokens, with some enterprise use cases supporting even longer inputs, and Anthropic reported meaningfully lower hallucination rates and improved instruction-following compared to Claude 2. Anthropic also emphasized Constitutional AI-based safety training and reduced unnecessary refusals in Claude 3, addressing a common criticism of earlier Claude versions being overly cautious. Claude 3 was made available through the Claude API, Claude.ai, Amazon Bedrock, and Google Cloud Vertex AI. It was superseded within the same generation by Claude 3.5 Sonnet in June 2024, which outperformed the original Claude 3 Opus on many benchmarks despite sitting at the mid-tier Sonnet level, and eventually by the Claude 4 family, but the Claude 3 tier naming convention (Haiku/Sonnet/Opus) has persisted across subsequent Anthropic releases.
Key Features
- Three-tier lineup: Haiku, Sonnet, and Opus, trading off speed, cost, and capability
- Native multimodal (vision) input across all three tiers at launch
- Context window support up to 200K tokens
- Claude 3 Opus benchmarked competitively with GPT-4 at release
- Reduced unnecessary refusals compared to Claude 2
- Trained using Anthropic's Constitutional AI safety approach
- Available via Claude API, Claude.ai, Amazon Bedrock, and Google Cloud Vertex AI
Use Cases
Alternatives
Frequently Asked Questions
From the Blog
Claude vs ChatGPT vs Gemini: Which Is Best?
A comprehensive guide to claude vs chatgpt vs gemini: which is best? — written for learners at every level.
Read More AI & TechnologyLarge Language Models (LLMs) Explained for Beginners
An LLM predicts the next piece of text, one token at a time — this guide explains how ChatGPT, Claude, and Gemini actually work.
Read More AI & TechnologyVibe Coding: How to Build Faster with AI Without Losing Control
AI coding tools have shifted from autocomplete to full code generation, multi-file refactoring, and autonomous debugging. This guide explains how to use tools like Copilot, Cursor, and Claude Code effectively — including the critical skill of reviewing AI-generated code before shipping it.
Read More AI & TechnologyBuilding Your First AI-Powered App with the Anthropic API
The fastest way to understand AI engineering is to build something real. This project- based guide walks you through building a writing assistant powered by Claude — from your first API call through streaming responses, a FastAPI backend, a simple frontend, and deployment.
Read More