Benchmarks show that Chinese AI models such as GLM-5.2, Kimi K3, and Qwen3.8-Max are now on par with or even better than major US models in some tasks, and often at significantly lower cost. The article highlights a developer who prefers these Chinese models.
Moonshot AI
Alibaba/Qwen
Cogent AI has released VR-1, a reasoning model post-trained specifically for cybersecurity, alongside IntrusionBench, a benchmark, and Cogent AI Harness, a governed runtime. The model is available only to vetted organizations through the Cogent Frontier Access Program and is aimed at large enterprises and government entities. Cogent reports VR-1 proves roughly twice as many attack paths at about a quarter of the cost compared to baselines like Kimi K3, Claude Opus 4.8, and GLM-5.2.
Cogent AI
Moonshot AI
Anthropic