Anthropic, Supported by Google, Launches Its Most Potent Chatbot Yet Amid Escalating Generative AI Competition

Anthropic made a revelation of Claude 3 on Monday, introducing a suite of artificial intelligence models that it claims are its fastest and most powerful to date. The new offerings, named Claude 3 Opus, Sonnet, and Haiku, mark a significant advancement for the company.

According to Anthropic, Claude 3 Opus, the most sophisticated of the new models, surpassed industry benchmarks such as undergraduate-level knowledge, graduate-level reasoning, and basic mathematics, outperforming OpenAI’s GPT-4 and Google’s Gemini Ultra.

For the first time, Anthropic is providing multimodal support with these models, allowing users to upload various types of unstructured data including photos, charts, and documents for analysis and answers.

Sonnet and Haiku, the other models introduced alongside Opus, offer a more compact and affordable alternative. While Sonnet and Opus are available in 159 countries starting immediately, Haiku will be released soon, according to Anthropic. The company refrained from disclosing the duration and cost of training Claude 3 but noted that companies like Airtable and Asana contributed to A/B testing the models.

Claude 3
Multimodal support introduced: Upload photos, charts, and documents for analysis and answers.

A year ago, Anthropic was recognized as a promising generative AI startup founded by former OpenAI research executives. Though it had completed Series A and B funding rounds, its initial product rollout lacked consumer access or significant attention.

Today, Anthropic stands as one of the leading AI startups, backed by major players like Google, Salesforce, and Amazon, with a product that competes directly with ChatGPT in both enterprise and consumer markets. Over the past year, the company secured five funding deals totaling approximately $7.3 billion.

The generative AI sector has experienced explosive growth, with a record $29.1 billion invested across nearly 700 deals in 2023, marking a more than 260% increase in deal value from the previous year, according to PitchBook. Despite concerns raised by academics and ethicists regarding bias propagation, generative AI has permeated various industries including education, healthcare, advertising, and more.

Between 60 and 80 individuals contributed to the core AI model development, while 120 to 150 individuals worked on its technical aspects, as noted by Anthropic co-founder Daniela Amodei in an interview with CNBC. Amodei further explained that for the AI model’s latest iteration, a team of 30 to 35 individuals directly worked on it, supported by approximately 150 individuals overall.

Claude 3’s advancements: Summarizes 150,000 words, better risk understanding, and nuanced prompt comprehension. (Credits: Unsplash)

Anthropic claims that Claude 3 can summarize up to approximately 150,000 words, equivalent to a substantial book such as “Moby Dick” or “Harry Potter and the Deathly Hallows,” doubling the capacity of its previous version. Users can input large datasets and request summaries in the form of memos, letters, or stories, while ChatGPT, in contrast, can handle around 3,000 words.

Amodei highlighted that Claude 3 demonstrates a better understanding of response risk compared to its predecessor, Claude 2, which sometimes erred on the side of caution, particularly concerning sensitive topics or trust and safety parameters.

Anthropic emphasizes that Claude 3 exhibits a more nuanced comprehension of prompts, reflecting advancements in generative AI capabilities. Multimodality, incorporating options like photo and video capabilities into generative AI, has emerged as a prominent use case in the industry.

Brad Lightcap, OpenAI’s COO, underscored the importance of multimodality, noting that the world encompasses various sensory modalities beyond text and code. However, the integration of multimodality and increasingly complex AI models presents potential risks. Google recently took its AI image generator offline after users identified historical inaccuracies and questionable responses.

Unlike Google’s model, Anthropic’s Claude 3 does not generate images but allows users to upload images and documents for analysis. Acknowledging that no model is flawless, Amodei emphasized Anthropic’s dedication to balancing capability and safety, albeit acknowledging occasional inaccuracies.

In the rapidly evolving landscape of AI, Anthropic’s Claude 3 represents a significant step forward, offering powerful multimodal capabilities while striving to mitigate potential risks inherent in advanced AI systems.

Sajda Parveen
Sajda Parveen
Sajda Praveen is a market expert. She has over 6 years of experience in the field and she shares her expertise with readers. You can reach out to her at [email protected]
Notify of
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x