Anthropic launches Claude 3.5 Sonnet to take on OpenAI's GPT-4o
What's the story
Anthropic, a rival to OpenAI, has unveiled its most advanced artificial intelligence model, Claude 3.5 Sonnet.
The company was founded by ex-OpenAI research executives and has received backing from industry behemoths like Google, Salesforce, and Amazon.
In the past year alone, Anthropic has secured approximately $7.3 billion through five funding deals.
The new AI model is part of the company's new Claude 3.5 family and is faster than its predecessor, Claude 3 Opus.
Model performance
Enhanced capabilities and new features
Anthropic claims that Claude 3.5 Sonnet demonstrates a "marked improvement in grasping nuance, humor, and complex instructions, and excels at writing high-quality content with a natural, relatable tone."
The AI model also has the ability to write, edit, and execute code.
Alongside this launch, Anthropic introduced "Artifacts," a feature that allows users to ask the Claude chatbot to generate text documents or code, which can then be opened in a dedicated window for real-time editing.
Performance
Claude 3.5 Sonnet sets new industry records
Claude 3.5 Sonnet has been reported to outperform the company's previous top-tier model, Claude 3 Opus, while operating at twice the speed.
The new model has broken industry records for graduate-level reasoning, undergraduate-level knowledge, and coding proficiency, even surpassing OpenAI's GPT-4o on many benchmarks.
However, competition remains tight as scores between ChatGPT, Claude, Gemini and Llama models tend to be within a few percentage points of each other on most tests.
Information
It also excels in visual input interpretation
Anthropic also claims that Claude 3.5 Sonnet is better at interpreting visual input. The new model can "accurately transcribe text from imperfect images," a feature that could attract customers in logistics, retail, and financial services who need to interpret data from charts and graphs.
Business plan
Anthropic expands enterprise offerings with team
Anthropic has recently introduced a business plan called Team after beta-testing with between 30 to 50 customers from various industries.
"So much of what we were hearing from enterprise businesses is people are kind of using Claude at the office already," Anthropic co-founder Daniela Amodei to CNBC.
The company also welcomed new team members last month, including Instagram co-founder Mike Krieger as chief product officer and OpenAI's former safety leader Jan Leike to lead new 'Superalignment' team.