On June 20, 2024, Anthropic released Claude 3.5 Sonnet—and it immediately became developers' favorite AI.

It beat GPT-4o on coding benchmarks despite being a "mid-tier" model. And it introduced Artifacts—a feature that let you see and interact with code Claude wrote in real-time.

What Made It Special

Coding performance: Beat GPT-4o and Claude 3 Opus on programming tasks Artifacts: Live preview of code, documents, and visualizations Speed: Faster than Opus while being more capable Cost: Same price as previous Sonnet (middle tier)

It was better and cheaper than Anthropic's own flagship.

Artifacts Changed Everything

What Artifacts did:

• Show code output live as Claude wrote it
• Let you modify and iterate visually
• Create interactive demos instantly
• Build complete mini-apps in the chat

It transformed Claude from chatbot to development environment.

The Coding Benchmarks

Claude 3.5 Sonnet scored:

• HumanEval: 92% (vs GPT-4o's 90%)
• Better at debugging: Faster at finding and fixing issues
• More complete code: Fewer missing pieces

Developers noticed immediately.

Where Are They Now?

Claude 3.5 Sonnet remains many developers' preferred coding AI. The combination of performance and Artifacts made it distinctly useful for building software.

June 20, 2024 was when An thropic proved that mid-tier models could beat flagship competitors with the right training and features.