“A new way of measuring AI models suggests China's threat to OpenAI and Anthropic may have been overblown”

Key points

• Artificial Analysis overhauled its AI ranking system to reflect real-world use.
• Under the new system, the gap between open-weight models and overall leader grew to 12 points.
• Publicly released weights performed far worse than the top overall model in new metrics.
• Usage figures suggest most AI demand may come from closed models like OpenAI and Anthropic.
• Analysts question whether open-weight AI can justify premium pricing amid the revised benchmarks.