Home ยป JetBrains Unveils Mellum: A Language Model for Code Writing to Boost Code Efficiency Faster than Traditional LLMs

JetBrains Unveils Mellum: A Language Model for Code Writing to Boost Code Efficiency Faster than Traditional LLMs

JetBrains camp offers an AI-assisted code writing service known as JetBrains AI Assistant, which processes on the server side. This service has been available since 2023, utilizing language models from various sources such as OpenAI and Google LLM.

Recently, JetBrains introduced their own language model called Mellum, specifically designed for code completion tasks. This model is more compact and efficient for code writing tasks, leading to faster code completion compared to other language models in the market.

Currently, the JetBrains AI Assistant service has switched to using Mellum for code completion (while other tasks still utilize other LLM models). It has resulted in a reduction of code completion waiting times by up to 2/3, an acceptance rate of 40%, a doubled amount of code completions shown to users, and clearer user feedback.

Mellum is currently compatible with languages such as Java, Kotlin, Python, Go, and PHP, with support for other languages planned for the future.

TLDR: JetBrains AI Assistant service now uses Mellum for code completion, leading to faster completion times and improved user feedback.

More Reading

Post navigation

Leave a Comment

Leave a Reply

Your email address will not be published.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

Prediction by Sam Altman: AI Superintelligence on the Horizon in ‘Just a Few Thousand Days’ Ahead

Insights on Artificial Intelligence Quandaries Unraveled by Google with Unconventional Solutions for Quick and Continuous Improvement

Unveiling GitHub Copilot’s AI Model Options: Featuring Claude 3.5, Gemini 1.5 Pro, and o1-mini