Skip to content

Pick your style: Denkbaar

Book a conversation

Articles

History

Claude 3.5 Sonnet: faster and smarter than Claude 3 Opus

·

Anthropic released Claude 3.5 Sonnet: better than Claude 3 Opus on many evaluations, twice as fast, and with Artifacts as a new window next to the conversation.

A maker works with a conversation and a design side by side.
With Artifacts, you see the result beside the conversation. Illustration: Denkbaar.

On 21 June 2024 Anthropic’s news page shows “Introducing Claude 3.5 Sonnet”. The new model is a mid-tier model, but according to Anthropic it does better than Claude 3 Opus, their largest model until then.

What was new

  • Better and faster. According to Anthropic, Claude 3.5 Sonnet outperforms Claude 3 Opus “on a wide range of evaluations”, with the speed and cost of Claude 3 Sonnet. It runs at twice the speed of Claude 3 Opus.
  • Price. On the API it costs $3 per million input tokens and $15 per million output tokens.
  • Context. The model reads up to 200,000 tokens at once.
  • Where. Free on Claude.ai and in the iOS app. Also on the Anthropic API, Amazon Bedrock and Google Cloud Vertex AI.
A medium and a large module lie beside test cards.
According to Anthropic, the mid-tier model surpassed the earlier top model on many tests. Illustration: Denkbaar.

Artifacts

At the same time Artifacts came to Claude.ai. Ask for code, a document or a website design, and the result appears in its own window next to the conversation. You see what you make while you talk about it.

A laptop shows a conversation beside a design panel.
A separate panel makes the created work visible at once. Illustration: Denkbaar.

Why this matters

A mid-tier model that overtakes the top model, at the mid-tier price. And a window where you see at once what you make. That second idea, first see something and only then build on, returns in many later tools.

Sources

  1. Anthropic: Introducing Claude 3.5 Sonnet

Topics: modelsclaudeanthropic