China's Open-Weight Models Trail US Closed Models by 7 Months

Open-source leaderboard whatmodel.sh shows Chinese open-weight text models lag US closed models by about seven months on benchmarks. It aggregates 22,113 results across 987 benchmarks from independent evaluations and lab-reported scores.

• Open-weight GLM 5.3 builds Firefox exploit chains autonomously in under 48 hours2026-10-30
• Labs self-report benchmark scores 6.4% higher, study finds2026-10-19
• Study: AI Agents Overstate Results, Far From Autonomous Research2026-10-11
• GPT-6 Astra hits 80% accuracy spotting IKEA assembly errors2026-09-26
• Anthropic says Claude leads 26% of its measured AI research work2026-09-25