The tweet was deleted by the author.
But we saved everything 🙂.
Artificial intelligence developer Alexandr Wang announced that Muse Spark 1.1 has surpassed Opus, Grok 4.5, and Gemini based on results from a new, challenging evaluation focused on finite model theory and theoretical computer science. According to Wang, the test was designed to assess the sophistication and reasoning capabilities of advanced AI models in complex computational domains. The comparative results suggest Muse Spark 1.1 stands out among its peers in theoretical computer science tasks.
Wang has previously said Muse Spark 1.1 is among the highest performing models for OpenClaw and HermesAgent applications, according to a recent statement. In an earlier update, he cited rapid progress on Muse Spark, Muse Image, 1.1, and its API, with further advancements expected, as reported here.