Tweet Claims Mythos 5.1 Leads on AECI Metric
Pseudonymous analyst @scaling01 shares a chart of model scores on the benchmark.
TLDR
@scaling01 posted that Mythos 5.1 reached 161.98 on the AECI benchmark. The tweet presents this as the highest mark shown and includes a chart comparing it to earlier Claude models such as Opus 5. The account describes itself as running LisanBench and frequently posts technical analysis of model performance. No first-party statement from the Mythos team appears in the packet. The post stands as the sole source for the reported figure and ranking.
Combined views
77.9K
3 Sources, first seen 29d ago