iFlytek Spark 3.0 Outperforms GPT-3.5, Setting New Benchmark for Large Language Models

carri-insights-iflytek-spark-3.0-outperforms-gpt-3.5,-setting-new-benchmark-for-large-language-models

iFlytek, a Chinese AI company, claims its large language model, Spark 3.0, outperforms OpenAI’s GPT-3.5 in Chinese language tasks and equals it in English. Trained on a dataset of 1.2 trillion words, Spark 3.0 can perform various tasks including text generation and code generation. Despite these claims, iFlytek’s shares fell 10% due to controversial content on its educational tablet. The company plans to launch Spark 4.0, benchmarked against GPT-4, in H1 2024.

Read More: https://www.gizmochina.com/2023/10/26/iflytek-spark-3-0-vs-openai-gpt/

Leave a Comment

Your email address will not be published. Required fields are marked *