Skip to content
Live
Synced·Added· 23 days ago65

Kimi K3.1 and Qwen 4 were discovered in blind tests in Code Arena.

Brief overview

Kimi K3.1 and Qwen 4 were tested anonymously in Code Arena programming battles, attracting attention. Kimi K3.1 scored 89.3% and 86.3% in HumanEval and SWE-bench tests respectively, while Qwen 4 achieved 89.4% in HumanEval.

Source: Synced · jiqizhixin.com