Independent Benchmark Puts Kimi K3 Third on Intelligence Index, Behind Fable 5 and GPT-5.6 Sol

Artificial Analysis has published independent test results for Kimi K3, the 2.8-trillion-parameter mixture-of-experts model Moonshot AI launched this week, placing it third on the firm's Intelligence Index behind Claude Fable 5 and GPT-5.6 Sol.

• 57 on the Artificial Analysis Intelligence Index, rank #3 overall, roughly level with Opus 4.8 and GPT-5.5
• Ahead of GLM-5.2 (51) and DeepSeek V4 Pro (44)
• Ranks #1 on AutomationBench-AA at 53%, and scores a 1,668 Elo on GDPval-AA v2
• Costs $0.94 per task on Intelligence Index evaluations, using 21% fewer output tokens than predecessor K2.6
• API priced at $3/$15 per million input/output tokens, with a 1M-token context window; open weights are promised by July 27

The result confirms Kimi K3 as a genuine frontier contender rather than just a cheap open alternative, though it still trails the top two closed models on raw intelligence. Readers can compare it directly against Fable 5, GPT-5.6 Sol and other flagships in the arena.