Artificial Analysis's latest rankings reveal that four frontier models were released in just eight days: Grok 4.5, GPT-5.6, Muse Spark 1.1, and Kimi K3, all entering the top tier.
There are now six model teams with intelligence scores above 50. Back in June, only OpenAI and Anthropic had crossed that threshold.
Kimi K3 scored 57, placing third — behind Claude Fable 5's 60 and GPT-5.6 Sol's 59, but ahead of Claude Opus 4.8's 56.
The price competition is even more striking. Under a unified benchmark and official pricing, Kimi K3 costs $0.94 per task — roughly half the cost of Opus 4.8. GPT-5.6 Sol, just one point off the top, comes in at $1.04 per task. Meanwhile, Claude Fable 5 costs $2.75.
Grok 4.5 scored 54 and costs just $0.31 per task. In a mere eight days, the cost of near-frontier models has dropped to about one-half to one-third of what they were.
https://twitter.com/artificialanlys/status/2078165665278730490?s=46