BREAKING
US AI Labs Sound Alarm on Chinese Models
0
%
cheaper (low)
0
%
cheaper (high)
Same Workload: $544 vs $4,811
GLM-5.2
544
Opus 4.8
4811
Cost Advantage, Little Quality Gap
GLM-5.2
China
●
Within ~1 point of Opus 4.8
●
Edges GPT-5.5 in some SWE-bench
US frontier
OpenAI/Anthropic
●
Pricier offerings
●
Cited security caution
Why Open Weights Worry Regulators
1
Weights released
↓
2
Copied and tuned
↓
3
Safeguards fail
↓
4
Misuse risk
Threat to Contain or Force to Manage
AI NEWS BLITZ
OpenAI and Anthropic executives are warning about cheap, powerful Chinese AI models.