Anthropic Unveils Claude 3: New AI Models Explained
Get the lowdown on Anthropic's latest AI offerings: Claude 3 Opus, Sonnet, and Haiku.
The AI competition continues to heat up, with Anthropic becoming the latest contender to upgrade its capabilities through the launch of Claude 3.
Claude 3 is the new large language model (LLM) developed by the AI startup Anthropic. This model now drives the company's Claude AI chatbot. There are actually three distinct models within the Claude 3 family: the top-tier Claude 3 Opus, the mid-range Claude 3 Sonnet, and the yet-to-be-released Claude 3 Haiku.
Anthropic boldly asserts that Claude 3 outperforms not only its predecessor, Claude 2.1, but also rivals like OpenAI's GPT-4 and Google's Gemini Ultra. So, what exactly is Claude 3, and how does it measure up to the competition? Let's dive in.
Understanding Claude 3: Opus, Sonnet, and Haiku
Claude 3 represents a new series of LLMs created by Anthropic, an AI company with significant backing (including a hefty $4 billion investment from Amazon). Anthropic is loosely aligned with the effective altruism movement, meaning it focuses on developing AI technologies for public benefit in a safe and ethical manner.
The three new LLMs fall under the Claude 3 banner. Anthropic's most sophisticated model is Claude 3 Opus, available exclusively to Claude Pro subscribers. The company's benchmarks suggest it possesses advanced reasoning skills, with a 50.4% score on graduate-level tests in comparison to GPT-4's score of 35.7%.
Claude 3 Sonnet is the next step down, offered to users without a paid Claude Pro subscription. While still described as highly capable, its graduate-level test score was 40.4% (again, surpassing GPT-4's 35.7%).
Finally, there's the unreleased Claude 3 Haiku. This model is positioned as the least sophisticated in terms of capabilities but is designed for lightning-fast responses.
Haiku's focus on speed makes it an attractive option for businesses. It could power customer-facing AI chatbots on websites, helping with customer queries, bookings, and other support tasks. It may also benefit businesses needing to quickly process and analyze large amounts of data.
Claude 3 Pricing: What's the Cost?
Accessing some Claude 3 models requires a Claude Pro subscription. This applies specifically to the most advanced option, Claude 3 Opus.
Consumer pricing for Claude 3 Opus is $20 per month. Businesses will deploy it on a custom basis, using a common token system. For Opus, the cost is $15 per million input tokens and $75 per million output tokens.
Claude 3 Sonnet is freely available as part of the open version of Claude AI accessible worldwide. Anthropic describes it as their "hard-working" model. Businesses can customize it at a cost of $3 per million input tokens and $15 per million output tokens.
Currently, pricing information for Claude 3 Haiku is only available on a token basis: $0.25 per million input tokens and $1.25 per million output tokens.
Claude 3 vs. GPT-4: Initial Impressions
We're in the process of extensively testing Claude 3 against major competitors like ChatGPT (powered by GPT-4). At this point, it's too early to offer a definitive comparison.
.jpg "GPT 3 Vs GPT4")
Anthropic has published performance tests showcasing the Claude 3 family against rivals like ChatGPT and Google's Gemini Ultra LLM. It's essential to note that these tests conducted by Anthropic itself likely utilize engineered prompts that highlight Claude 3's strengths.
This isn't to diminish the potential of Claude 3, which appears promising. However, independent testing is crucial to determine how it truly stacks up against the competition.
Ready to Try Claude 3? Here's How
If you're curious to experience Claude 3, the mid-tier Claude 3 Sonnet model is now live within Claude AI in many regions.
Using Claude 3 is simple:
Visit the Anthropic website.
Click the "Try Claude" button.
Sign in with your Google account (or create one).
You'll be in the Claude AI chat window, ready to test Anthropic's AI. If you're impressed, consider whether the extra power of Opus is worth the upgrade.