Opensource Competitor Of GPT-4 Is Here

In a new development in the field of generative AI, a new language model called Orca has emerged as a formidable contender to ChatGPT, demonstrating exceptional performance in zero-shot tasks across a range of benchmarks. 

Equipped with an impressive 13 billion parameters, Orca has surpassed the competition in various evaluations, including the rigorous Big Bench Hard (BBH) challenge, as well as esteemed academic exams like SAT, LSAT, GRE, and GMAT. 

(As per the presented paper, Orca, with 13 billion parameters, outperforms a wide range of foundation models including OpenAI ChatGPT as evaluated by GPT-4 in the Vicuna evaluation set.)

Orca's stunning achievement in the BBH benchmark is not to be underestimated, as it has outperformed conventional state-of-the-art instruction-tuned models by an astounding margin of over 113%. Similarly, in the AGIEval benchmark, Orca has showcased its superiority by delivering an impressive 42% improvement over the previous standard. 

What sets Orca apart from its counterparts is its ability to progressively learn from intricate explanation traces, including step-by-step thought processes, derived from the highly advanced GPT-4 model. This novel approach to learning has proven to be a game-changer, unlocking previously untapped capabilities and substantially enhancing Orca's skills. 

To comprehensively evaluate Orca's diverse abilities, a team of researchers conducted an extensive battery of tests covering writing, comprehension, analytical skills, mathematical prowess, logical reasoning, and safety considerations.