Anthropic Unveils Claude 3 AI With Expanded Abilities, Claiming New Benchmarks Over Rivals
-
Anthropic released Claude 3, the latest version of its AI chatbot, claiming it has "human-like understanding" and outperforms rivals like ChatGPT and Google's Gemini on benchmarks.
-
The Claude 3 family contains 3 models - Haiku, Sonnet, and Opus - with Opus being the most capable. They have a 200,000 token context window.
-
Compared to previous versions, Claude 3 is more accurate, can follow complex instructions better, and is less likely to refuse to answer harmless questions.
-
Anthropic said Claude 3 shows less bias and presents negligible potential for catastrophic risk based on internal testing.
-
The company plans frequent updates to Claude 3 over the next few months as it continues pushing the boundaries of AI capabilities.