Anthropic just dropped Claude Haiku 3.5 and gave the chatbot a huge upgrade — here's what's new
Plus, wait until you hear what's in beta.
Anthropic, a leader in the AI space, today announced a major upgrade to its Claude models with an update to its already impressive Claude 3.5 Sonnet and a brand new offering, Claude 3.5 Haiku.
In addition to these upgrades, Anthropic is rolling out an exciting new feature in beta, which allows AI models to interact with computers, essentially mimicking human interactions — an advancement poised to redefine AI capabilities.
The Claude 3.5 Sonnet builds on the capabilities of previous models, which can only mean a boost to performance in both comprehension and content generation. Sonnet’s refinement is sure to be an exciting improvement for the already user-friendly model.
Additionally, Claude 3.5 Haiku brings a fresh approach, which Anthropic boasts is a more concise natural language model that’s “three times faster" than its peers.
The model is designed to optimize performance in applications where quick, short-form responses are crucial. I’m looking forward to trying what seems to be an AI model tailor-made for tasks that require brief yet effective answers such as microcontent creation.
How the new Sonnet 3.5 stacks up?
Claude 3.5 Haiku is the next generation of our fastest model.Haiku now outperforms many state-of-the-art models on coding tasks—including the original Claude 3.5 Sonnet and GPT-4o—at the same cost as before.The new Claude 3.5 Haiku will be released later this month. pic.twitter.com/M38d8nleL7October 22, 2024
Anthropic published new benchmark results in its testing of both the updated Claude Sonnet 3.5 and the long-awaited Haiku 3.5.
Sonnet now comfortably outperforms GPT-4o from OpenAI and Gemini 1.5 Pro from Google on graduate-level reasoning tasks, coding and visual analysis.
Sign up to get the BEST of Tom's Guide direct to your inbox.
Get instant access to breaking news, the hottest reviews, great deals and helpful tips.
Haiku, when compared to similar-sized models from Google and OpenAI — Gemini Flash 1.5 and GPT-4o mini — is better at coding, agent-like behavior and text reasoning.
Anthropic wrote on X that: "Haiku now outperforms many state-of-the-art models on coding tasks—including the original Claude 3.5 Sonnet and GPT-4o—at the same cost as before."
Looking ahead with computer use capability
But perhaps the most exciting bit of news from Anthropic today is Anthropic’s new beta feature: computer use capability. This is sure to turn some industry heads as the update enables Claude models to interact with computers in a manner that simulates human behavior.
From viewing screens and moving a cursor to clicking and typing, the new AI opens a wealth of possibilities for how AI can assist. Essentially, the hope is that Claude can take the reins when it comes to performing everyday computer tasks.
While still in beta, the computer use feature promises to make AI more versatile, especially in handling complex workflows such as filling out forms and navigating websites.
Anthropic’s updates position Claude to compete more aggressively with some of its biggest competitors. As these innovations roll out, the AI community will be watching closely to see how these models are used in real-world applications.
More from Tom's Guide
- Meet Mochi-1 — the latest free and open-source AI video model
- Gemini vs. ChatGPT vs. Meta AI — which AI image generator wins with food?
- Ideogram drops new canvas editor — I got early access and it's a game-changer