American start-up AI company Anthropic has launched its latest generation of proprietary AI models called Claude 3. These models are set to rival industry giants such as Google’s Gemini and OpenAI’s GPT-4. The company claims the Claude 3 family of LLMs (Large Language Models) will “set new industry benchmarks” for generative AI.
“Today, we’re announcing the Claude 3 model family, which sets new industry benchmarks across a wide range of cognitive tasks”, the company stated in a blog release.
See Related: OpenAI Launches ChatGPT Plus Subscription In India; Includes GPT-4
The blog also confirmed that Claude 3 will come in 3 versions: Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus. According to Anthropic, “each successive model offers increasingly powerful performance, allowing users to select the optimal balance of intelligence, speed, and cost for their specific application”.
Anthropic has released several benchmark test results that show Claude 3 Opus outperforming its competition. For example, Opus delivers superior results than both GPT-4 and Gemini 1.0 in undergraduate level expert knowledge (MMLU), basic mathematics (GSM8K), mixed evaluation (Big-Bench-Hard), and more. It also claims Claude is capable of “near-human levels of comprehension”.
Both Pus and Sonnet are available on claude.AI and on Claude API in 159 countries. Claude Haiku will be released sometime in the future. The models will offer a 200K context window at launch, with the models capable of exceeding 1 million tokens.