Article complete

Meta Releases Code Llama 70B

This model achieves an impressive 67.8 on HumanEval.

A stylized llama with the Meta logo overlayed, surrounded by text and digital elements.

Key takeaways

  • Meta launches Code Llama 70B, enhancing open-source code generation capabilities.
  • New model achieves 67.8 on HumanEval, ranking among the best open models.
  • Offers three free versions for diverse research and commercial applications.
  • Mark Zuckerberg emphasizes importance of coding in AI model advancements.
  • Code Llama builds on Llama 2, continuing Meta's commitment to open-source innovation.

Takeaways drafted with AI from the article text.

One notable addition to the suite is CodeLlama-70B-Instruct, a finely tuned version specifically designed for instruction-based tasks. This model achieves an impressive 67.8 on HumanEval, solidifying its status as one of the highest-performing open models currently available. 

Code Llama 70B provides three free versions for research and commercial use: foundational code (CodeLlama - 70B), Python specialisation (CodeLlama - 70B - Python), and fine-tuned for natural language instructions (Code Llama - 70B - Instruct 70B).

“We're open sourcing a new and improved Code Llama, including a larger 70B parameter model. Writing and editing code has emerged as one of the most important uses of AI models today,” said Meta chief, Mark Zuckerberg. 

“The ability to code has also proven to be important for AI models to process information in other domains more rigorously and logically. I'm proud of the progress here, and looking forward to including these advances in Llama 3 and future models as well,” he added. 

Meta released its code generation model called Code Llama, last August, which generates code based on both code and natural language prompts. The best part is that just like Llama 2, Code Llama is open source and also available for commercial use. 

Code Llama is built upon the foundation of Llama 2, which has been fine-tuned with specialised code-related datasets. The company announced four versions of Code Llama — Code Llama, Code Llama Instruct, Code Llama Python, and Unnatural Code Llama, each with varying capacities: 7B, 13B, and 34B parameters.

Published

Spotted an error? Tell our editors. Read our editorial standards and corrections.

About the author

Tech Journalist

Siddharth is a media graduate who loves to explore tech through journalism and putting forward ideas worth pondering about in the era of artificial intelligence.

All stories
Belamy · Daily

The day’s AI news, before your first meeting.