Blogger's internal benchmarks: Chinese AI models pricier than US models quality-adjusted

kevinnbass · x · 2026-09-08

The author says internal benchmarking shows Chinese models aren't that great and cost more than American models on a quality-adjusted basis; US AI is in better shape than the hype suggests, though that may not last. He also torches VC discourse on X: VCs rarely go beyond a few prompts in a harness, are consistently wrong, and may fund fake companies — never listen to them.

Original post →

More from Models

Models channel →