Small Qodo model outperforms giants OpenAI and Salesforce
Qodo company, previously known as Codium, has introduced a revolutionary open model Qodo-Embed-1-1.5B for working with program code, which has surpassed solutions from technology giants OpenAI and Salesforce, despite its more compact size and high efficiency.
The new model with 1.5 billion parameters sets a new standard in the field of search, extraction, and understanding of code, which is especially important for corporate development teams managing large-scale and complex codebases.
“Corporate software can contain tens, if not hundreds of millions of lines of code. Code generation alone is not enough — you need to ensure its high quality, correct operation, and integration with the rest of the system,” explains Itamar Friedman, CEO and co-founder of Qodo.
According to Friedman, context plays a key role in modern software development using AI models. “To extract the right context from a really large codebase, you need to use a specific search mechanism,” he notes.
Unlike traditional artificial intelligence solutions focused on code generation, Qodo-Embed-1-1.5B offers a more comprehensive approach. The model is particularly effective for large organizations where software projects span millions of lines of code distributed across multiple teams, repositories, and programming languages.
The release of this model as open source represents a significant step forward in the development of AI tools for software development, offering a more accurate and efficient solution for searching and analyzing code in large-scale corporate systems.
