New Intelligence Index Ranks Frontier AI Models

Artificial Analysis has developed an "Intelligence Index" to quantify the capabilities of frontier AI models. This index is a weighted average of nine evaluations, with a strong emphasis on agentic tasks. While closed-source models currently lead in all three of the index's categories, the comparison is limited by the lack of transparency regarding their size and architecture. The top-performing open-weight model, GLM-5.2, is a fully documented 753B mixture of experts. AI

IMPACT Provides a new quantitative framework for comparing AI model capabilities, highlighting the lead of closed-source models and the performance of open-weight alternatives.

RANK_REASON The cluster describes a new benchmarking methodology for AI models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — fosstodon.org →

• Artificial Analysis
• GLM-5.2

• model release
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →