LLM & Text Generation4 min reading time

The Biggest Benchmark in AI Just Updated, Dealing a Major Blow to a Chinese Contender

Gizmodo
Read the full article
Moonshot AI's Kimi K3.1 Pro initially ranked 4th on the Artificial Analysis Intelligence Index v4 but dropped to 9th after the benchmark updated to v5, which includes more private test questions to prevent overfitting. Google’s Gemini 4 Carbon now leads the new index, highlighting a wider performance gap between open-weight and proprietary models.

More on this story


More in LLM & Text Generation