Moonshot AI's Kimi K3 model has demonstrated that China is now competitive with leading AI models, according to Artificial Analysis. Kimi K3 achieved a score of 57 on the Intelligence Index, closely trailing Claude Fable 5 (60) and GPT-5.6 Sol (59). Notably, Kimi K3 offers this near-frontier capability at a significantly lower cost, priced at $0.94 per task compared to Sol's $1.04, and is substantially cheaper than DeepSeek V4 Pro. This development suggests a narrowing gap in both AI capability and cost, with Moonshot AI planning to release the model's weights on July 27. AI
IMPACT Narrows the capability and cost gap between leading AI models, potentially influencing enterprise adoption strategies.
RANK_REASON Model release from a frontier lab (Moonshot AI) with benchmark comparisons. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Artificial Analysis
- Caspar Bannink
- China
- Claude Fable 5
- DeepSeek V4 Pro
- GPT-5.6 Sol
- HomeScout
- Kimi k3
- Moonshot AI
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
