AI Model Leaderboard

Agentic work ranking, Model families, sorted by Overall
Rank Model Provider Weights Overall Coding Agents Reasoning Math Knowledge Cost
129 Kimi K2 Instruct Moonshot AI Open 46.1 46.6 46.4 44.2 –No results 42.5 $0.88
124 Kimi K2 Thinking Moonshot AI Open 47.9 48.0 47.6 49.3 50.1 44.8 $1.08
89 Kimi K2.5 Moonshot AI Open 53.1 54.4 52.5 53.4 53.5 51.9 $1.20
70 Kimi K2.6 Moonshot AI Open 56.0 56.7 55.1 58.4 56.7 51.9 $1.71
65 Kimi K2.7 Code Moonshot AI Open 56.6 55.3 57.2 56.1 56.5 52.0 $1.71
35 Kimi K3 Moonshot AI Open 59.9 59.5 59.4 62.1 60.2 58.3 $6.00
15 Kimi K3.1 Pro Moonshot AI Open 62.4 61.2 62.6 63.9 62.3 60.5 $8.00