The article frames Chinese open-weight models as leading US counterparts on downloads and cited capability scores, but the supplied commentary alone does not independently verify the comparison.
Limits of the evidence: This is commentary based on the author's stated tracking and cited benchmark results; the supplied text does not provide underlying datasets or independent corroboration, and the excerpt ends mid-sentence.
What to watch
Assessment revised Oct 4, 2026
The current balance of power in open models
InterconnectsSep 21, 2026First report
First assessment, from Interconnects
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.