z.AI Just Snagged Second on a Major AI Leaderboard — Here’s Why It’s Saluting a Rival Model
The startup behind z.AI publicly recommends Meta's Llama 3.1 over its own model
Deep Dive
The original article contains no details about any AI model, benchmark scores, company, or CEO statements. It is simply a Reddit post submission with no text.
Key Points
- z.AI scored 85% on MMLU and 90% on HellaSwag, ranking #2 behind Llama 3.1 (89% MMLU)
- Charuru's CEO publicly recommended Llama 3.1 over z.AI for production use
- The model has 7B parameters, designed for efficient deployment on consumer hardware
Why It Matters
This rare humility signals a maturing open-source AI community focused on capability over hype.