Anthropic has released Claude Sonnet 3.5, which is more powerful, faster and cheaper than the larger Claude 3 Opus model.
When Anthropic Claude They were added to the brand new Model 3 family in March and were available in three variants: Haiku, Sonnet and Opus, each of which had larger, more powerful and dearer tokens.
Claude Sonnet 3.5 is significantly more intelligent than its larger predecessor and has significantly improved image processing and coding capabilities.
It's also much faster and cheaper. Anthropic says that inference with Claude Sonnet 3.5 is twice as fast as with Claude Opus 3, is five times cheaper per token, and in addition has a context window of 200,000.
In just three months, Claude Opus 3 has turn into obsolete, and Anthropic says we are able to expect updated 3.5 versions of Haiku and Opus “soon.”
Anthropic provides the model freed from charge on its Claude.ai Chat interface and iOS app. If you enroll for a paid account, you’re going to get higher rate limits and API access.
Benchmark results for Claude Sonnet 3.5
While Claude Sonnet 3.5 can't search the net or generate images, its improved image processing, mathematics, reasoning, and coding capabilities outperform industry leaders GPT-4o and Gemini Pro 1.5 in a spread of benchmarks.
The leads to the realm of visual mathematical considering and programming stand out here, and users are particularly excited concerning the improved programming skills.
I’m really impressed by Claude 3.5 Sonnet programming skills.
I created this visualization of chaos using 40 triple pendulums, each with barely different initial conditions, in about 5 minutes after just a few iterations! Without Claude, it will have easily taken me hours. pic.twitter.com/RhCKhFwUyu
— Luis Batalha 🇵🇹🇺🇸 (@luismbat) June 22, 2024
Artifacts
The Artifacts feature is an exciting addition to Claude's web chat interface. ChatGPT generates code for you, but then you will have to repeat and paste it right into a development environment to try it out.
Claude now has an extra window that opens next to the chat interface where you’ll be able to see a real-time preview of the code. Changes are immediately reflected within the Artifacts window.
Anthropic says that Artifacts will soon support teams and enable collaboration on projects. Let's hope ChatGPT gets its own version of Artifacts soon.
Anthropic said it has subjected Claude 3.5 Sonnet to rigorous security testing and has also submitted it to the UK Artificial Intelligence Safety Institute (UK AISI) for pre-deployment security assessment.
Its internal security assessment, published within the sample cardclassified “Claude 3.5 Sonnet as an AI Assurance Level 2 (ASL-2) model, indicating that it poses no risk of catastrophic harm.”
Anthropic says that along with updated versions of the Haiku and Opus models, modalities, storage capability and other enterprise integration features will soon be added.
