Claude 3 (Opus / Sonnet / Haiku)
AI Model
Language
Historical
The March 2024 family that first beat GPT-4 on public benchmarks.
You're looking at an AI model record: who built it, what it can do, and the hardware and data behind it.
Claude 3 established Anthropic's three-tier naming — Opus, Sonnet, Haiku — and its Opus tier was the first model to clearly top GPT-4 on standard benchmarks, with vision support and 200K context. It moved Anthropic from safety-lab curiosity to full commercial frontier competitor.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Developed by
Based on
Succeeded by
Competes with
Sources
Record reviewed as of 2026-01-15 — see Methodology.