Ryan Shrout @ryanshrout · 03 Sep 2026
Meta may have just had its next Llama moment. Muse Spark 1.3 lands SECOND on @Signal_65 PINNACLE, 38% fewer weighted errors than Claude Opus 5, and a correct task costs $0.85 against $1.36 for Opus and $1.22 for GPT-5.6 Sol. Both of those just fell off the cost frontier. It
756Views
5Likes
2Reposts
1Replies
0Quotes
1Bookmarks
Is that a lot?
0.37×vs this author's median2 038 views is typical
14Percentile for this authorof 7 recent posts
0.37×vs 10K–100K median2 043 views is typical
3.88%Reachviews ÷ followers
1.06%Engagement rateof viewers reacted