Anthropic Launches Claude 3.5 Sonnet, Its Newest AI Model

Just like other products, services and technology, generative artificial intelligence (GenAI) platforms, models and tools also need to be updated regularly.

The idea is that they should offer the latest capabilities, considering new technologies and trends in the industry.

Just over three months ago, we reported on artificial intelligence (AI) technology company Anthropic releasing its Claude 3 model family, which “sets new industry benchmarks across a wide range of cognitive tasks.”

This family consisted of three state-of-the-art generative models named Claude 3 Haiku, Claude 3 Sonnet and Claude 3 Opus (in ascending order of capability).

Now, the AI company has announced the next iteration of its Claude 3 family – Claude 3.5 Sonnet – which comes in with enhanced capabilities.

So, what did Anthropic reveal about its newest AI model? Let’s explore!
 

What Does Anthropic’s Claude 3.5 Sonnet Offer?

 

• “Introducing Claude 3.5 Sonnet — our most intelligent model yet,” read Anthropic’s post on X announcing the release of its newest model.

• The post continued by saying, “This is the first release in our 3.5 model family. Sonnet now outperforms competitor models on key evaluations, at twice the speed of Claude 3 Opus and one-fifth the cost.”

• Anthropic also published a news article on its website providing details about the model, along with a video exhibiting its capabilities.

• This included mentioning that the model would be available for free on Claude.ai and the Claude iOS app, as well as through the Anthropic API, Amazon Bedrock and Google Cloud’s Vertex AI.

• Furthermore, it clarified that Claude Pro and Team plan subscribers will get to use the model with “significantly higher rate limits”.

• As for pricing, the model costs $3 per million input tokens and $15 per million output tokens, with a 200K token context window.

• Ahead of this, Claude 3.5 Haiku and Claude 3.5 Opus will be released later in the year.

 

What Improvements Does Claude Sonnet 3.5 Bring?

 

• As per the release, Claude 3.5 Sonnet outperforms competitors’ models across various evaluations, with the speed and cost of its mid-level Claude 3 Sonnet model.

• This also includes better results against Claude’s own Opus, operating at twice the speed and solving 64% of problems in comparison to Claude 3 Opus’ 38%.

• Furthermore, it shows improvements in grasping nuance, humor and complex instructions, while writing high-quality content with a natural and relatable tone.

• It even offers improvements when the model’s ability to fix bugs or add functionalities to an open-source codebase, “given a natural language description of the desired improvement.”

• The model can even independently write, translate, edit and execute code with sophisticated reasoning and troubleshooting capabilities, when provided with the right tools.

• Claude 3.5 Sonnet offers state-of-the-art vision capabilities for tasks that require visual reasoning.

• Anthropic also brings in a new feature – Artifacts – which, when asked to generate content like code snippets, text documents or website designs, displays a separate dedicated window enabling a dynamic workspace where users can see, edit and build upon Claude’s creations in real-time.

• “Claude 3.5 Sonnet sets new industry benchmarks for graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding proficiency (HumanEval).”

• Additionally, Claude builds on its commitment to safety and privacy.

 

What Did Anthropic Say About The Safety Measures Adopted For Claude Sonnet 3.5?

 

• “Our models are subjected to rigorous testing and have been trained to reduce misuse. Despite Claude 3.5 Sonnet’s leap in intelligence, our red teaming assessments have concluded that Claude 3.5 Sonnet remains at ASL-2.”

• “As part of our commitment to safety and transparency, we’ve engaged with external experts to test and refine the safety mechanisms within this latest model.

• “We recently provided Claude 3.5 Sonnet to the UK’s Artificial Intelligence Safety Institute (UK AISI) for pre-deployment safety evaluation. The UK AISI completed tests of 3.5 Sonnet and shared their results with the US AI Safety Institute (US AISI).”

• “We have integrated policy feedback from outside subject matter experts to ensure that our evaluations are robust and take into account new trends in abuse.”

• “This engagement has helped our teams scale up our ability to evaluate 3.5 Sonnet against various types of misuse. For example, we used feedback from child safety experts at Thorn to update our classifiers and fine-tune our models.”

• “One of the core constitutional principles that guides our AI model development is privacy. We do not train our generative models on user-submitted data unless a user gives us explicit permission to do so. To date we have not used any customer or user-submitted data to train our generative models.”

What do you think of Anthropic’s Claude Sonnet 3.5? Do you think the move will help it increase its market share?

Let us know in the comments below!