Claude 3 is expected to better follow multi-step instructions, produce structured output, and converse in languages other than English. It also has an expanded context window, allowing it to better grasp the narrative flow of data and generate more contextually rich responses. However, it isn't immune from issues like bias and hallucinations, and can't search the web. Anthropic plans to release enhancements to the Claude 3 model family over the next few months. The models are available on the web and via Anthropic’s dev console and API, Amazon’s Bedrock platform, and Google’s Vertex AI.
Key takeaways:
- AI startup Anthropic has announced the latest version of its GenAI tech, Claude 3, which includes models Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus. The company claims Claude 3 outperforms AI chatbot OpenAI’s GPT-4 in terms of performance.
- Claude 3 is Anthropic’s first multimodal GenAI, meaning it can analyze text as well as images. However, it has limitations in identifying people and processing low-quality images or tasks involving spatial reasoning.
- Despite its improvements, Claude 3 still has issues common to GenAI models, such as bias and hallucinations. It also cannot search the web and is less fluent in certain "low-resource" languages compared to English.
- Anthropic plans to release enhancements to the Claude 3 model family over the next few months. The company is also aiming to raise as much as $5 billion over the next 12 months to remain competitive with OpenAI.