Right in the middle of OpenAI’s CEO ouster chaos, rival company Anthropic has released its latest AI model, Claude 2.1 with remarkable upgrades across the board. This not only includes a much bigger context window but also fewer model hallucinations and improved comprehension.
Industry Leading Context Window
OpenAI recently released GPT 4 Turbo with a context window of a whopping 128,000 tokens, but Anthropic has taken it a step further with Claude 2.1. The new AI model can take in as much as 200,000 tokens, putting it to industry-leading standards. This means it can work with up to 150,000 words, or 500 pages of material, letting people work with comprehensive and technical documentation.
2x Less Hallucination Rates
Hallucination in AI models is when the chatbot produces false information. Claude 2.1 successfully reduces hallucination rates by 2x, making it far more reliable than Claude 2.0. This is tested by making the AI model go through a large set of complex and factual questions and probing for errors.
This update also replaces false information with admissions of uncertainty. For instance, instead of saying “The fifth most populous city in Bolivia is Montero”, that chatbot would say: “I’m not sure what the fifth most populous city in Bolivia is.”
Improved Comprehension and Summarization
Claude 2.1 also comes with enhanced comprehension and summarization, particularly in handling lengthy and intricate documents that require a high level of accuracy, such as legal documents, financial reports, and technical specifications.
According to Anthropic’s blog post: “Claude 2.1 demonstrated a 30% reduction in incorrect answers and a 3-4x lower rate of mistakenly concluding a document supports a particular claim.”
Tool Use API
Anthropic has also added a heavily requested feature, tool use, that enables Claude to integrate with users’ existing processes, products, and APIs. This should make Claude more useful across users’ day-to-day operations, though the feature is only in beta as of now.
Thanks to tool use, Claude can coordinate actions across functions or APIs as defined by developers, perform searches across web sources, and retrieve information from private knowledge bases.
Last but not least, Anthropic has also made changes to its pricing models to “improve cost efficiency for our customers across models.”
“Claude 2.1 is available now in our API and is also powering our chat interface at claude.ai for both the free and Pro tiers. Usage of the 200K token context window is reserved for Claude Pro users, who can now upload larger files than ever before.”