BREAKING
Chinese AI Models Rival Anthropic Mythos
0
%
GLM-5.2 F1
0
%
Claude Code F1
0
$
per vuln found
GLM-5.2 Climbs on Terminal-Bench 2.1
GLM-5.1
63.5
GLM-5.2
81
Opus 4.8
85
GLM-5.2 vs 360's Tulongfeng
GLM-5.2
MIT open-weight
●
Single MoE model
●
~744-750B params
●
1M token context
Tulongfeng
360 Security
●
Multi-agent swarm
●
3,432 vulns found
●
105 confirmed
Open Weights Fill US Restriction Gap
US-China AI Gap Narrows in Cyber
AI NEWS BLITZ
Two Chinese AI offerings are reportedly matching Anthropic's Mythos on key cyber tasks.