Anthropic, co-founded by ex-OpenAI experts, drops an updated version of its AI model via Claude Instant 1.2 API. This release brings major enhancements in math, coding, reasoning, and safety. In tests, it scored 58.7% in coding and 86.7% in math, outshining its predecessor.
Claude Instant 1.2 generates more structured, formatted responses and rocks improved quote extraction, multilingual skills, and question answering. Less hallucination, increased resistance to jailbreaking, and a 100,000-token context window make it solid.
Anthropic's aim is a next-gen self-teaching AI algorithm. Claude Instant isn't that, but it's a competitor to OpenAI's offerings and other startups. It's here to shine in the realm of text generation and make its mark in the AI landscape.
Anthropic, launched in 2021 by ex-OpenAI VP Dario Amodei, has raised $1.45B and counts thousands of customers like Quora and DuckDuckGo. But it's aiming high, needing $5B in two years to create its envisioned chatbot. Exciting times ahead in the AI evolution!