Advertising

Anthropic’s AI chatbot surpasses OpenAI’s GPT-4 models

AI Startup Anthropic Unveils New GenAI Chatbot, Claude 3

Anthropic, an AI startup backed by Google and venture capital funding, has announced the release of its latest GenAI technology, Claude 3. The company claims that Claude 3 surpasses OpenAI’s GPT-4 models in terms of performance.

Claude 3 is a family of models that includes Haiku, Sonnet, and Opus, with Opus being the most powerful. Anthropic claims that all versions of Claude 3 show increased capabilities in analysis and forecasting compared to models like ChatGPT and GPT-4.

One notable feature of Claude 3 is its ability to analyze both text and images. Similar to some flavors of GPT-4 and Gemini, Claude 3 can process various visual elements such as photos, charts, graphs, and technical diagrams from different document types. It can also analyze multiple images in a single request, allowing for comparisons and contrasts between images.

However, there are limitations to Claude 3’s image processing capabilities. Anthropic has disabled the models from identifying people, likely due to ethical and legal concerns. The company also acknowledges that Claude 3 may make mistakes with low-quality images and struggles with tasks involving spatial reasoning and object counting. Additionally, Claude 3 is strictly an image-analyzing model and does not generate artwork.

According to Anthropic, customers can expect improvements in multi-step instructions, structured output formats, conversational abilities in languages other than English, and a more nuanced understanding of requests compared to previous models. Anthropic also plans to introduce features that will allow the models to cite the sources of their answers.

Claude 3’s enhancements are attributed to its expanded context window, which refers to the input data the model considers before generating output. Anthropic states that Claude 3 initially supports a 200,000-token context window, equivalent to about 150,000 words, with select customers having access to a 1-million-token context window (~700,000 words).

Despite its advancements, Claude 3 is not immune to the issues faced by other GenAI models, such as bias and hallucinations. Anthropic acknowledges these challenges in a technical whitepaper. It’s worth noting that Claude 3 cannot search the web and can only answer questions using data from before August 2023. Additionally, while Claude 3 is multilingual, it may not be as fluent in certain low-resource languages compared to English.

Anthropic plans to provide frequent updates to Claude 3 in the coming months. The company believes that model intelligence has not reached its limits and intends to enhance the Claude 3 model family.

Opus and Sonnet models are currently available on the web, Anthropic’s dev console and API, Amazon’s Bedrock platform, and Google’s Vertex AI. Haiku will be released later this year.

The pricing breakdown for the models is as follows:

– Opus: $15 per million input tokens, $75 per million output tokens
– Sonnet: $3 per million input tokens, $15 per million output tokens
– Haiku: $0.25 per million input tokens, $1.25 per million output tokens

Anthropic’s ultimate goal is to develop an AI self-teaching algorithm for building virtual assistants capable of various tasks such as answering emails, performing research, and generating art and books. The company plans to enhance Claude 3’s capabilities by allowing it to interact with other systems, code interactively, and deliver advanced agentic capabilities.

Anthropic’s approach to training GenAI, called “constitutional AI,” aims to align AI with human intentions by providing a simple set of guiding principles. For example, Anthropic has added a principle to Claude 3 that instructs the models to be understanding and accessible to people with disabilities.

With its ambitious plans and significant funding, Anthropic is poised to compete with OpenAI in the AI space. The company aims to raise up to $5 billion over the next 12 months, with strong support from Google, Amazon, and other backers. Anthropic’s commitment to continuous improvement suggests that we can expect further advancements from the company in the future.