Kimi K3 Tops Arena.ai's Frontend Code Leaderboard, Ahead of Claude Fable 5 and GPT-5.6 Sol

Moonshot AI's Kimi K3 has taken the top spot on Arena.ai's Frontend Code leaderboard, a blind human-evaluated benchmark for front-end coding tasks, days after the model's July 16 launch. That's a separate result from the Artificial Analysis Intelligence Index, where Kimi K3 placed third.

• Kimi K3 scored 1,679, ahead of Claude Fable 5 (1,631), GPT-5.6 Sol (1,618), GLM-5.2 (1,587), Claude Opus 4.8 (1,562), and Grok 4.5 (1,558)
• K3 ranked first in six of seven front-end categories (branding, reference-based design, data/analytics tools, consumer products, simulations, content-creation tools); Claude Fable 5 held only the Gaming category
• Predecessor Kimi K2.6 ranked 18th on the same leaderboard, so K3's debut is a 17-place jump
• K3 runs on a 2.8 trillion-parameter mixture-of-experts architecture with a 1M-token context window; full open weights are due July 27

Unlike traditional coding tests that focus on isolated functions or algorithms, the benchmark evaluates planning, debugging, tool use, interface design, and full project execution through blind human evaluations - TechStartups

Compare front-end coding output yourself in the coding arena.