Muse Spark 1.1 Nears Top Models in Benchmarks

alexandr_wang · x · 2026-07-12

Muse Spark 1.1 was benchmarked against several frontier models, showing it rivals Grok 4.5 on multiple high-signal evaluations and scores 51 on the Artificial Analysis Intelligence Index.

The post notes an 8-point jump from Muse Spark 1.0 over three months, driven mainly by science reasoning, coding, and knowledge. However, it still trails leaders in agentic knowledge work. Meta also granted the author early access for testing prior to the official release.

Related event: Meta Muse Spark 1.1 Shines Across Multiple Benchmarks(13 posts)→

Original post →

More from Models

Models channel →