Meta’s chief AI officer and highest-paid employee Alexandr Wang took a fresh swipe at Google after Meta’s latest AI model moved close to the top of an industry benchmark.
Wang reacted to the performance of Meta’s Muse Spark 1.3 on the Artificial Analysis Intelligence Index. In a post on X, formerly known as Twitter, he wrote: “I really hate to say it, but… Gemini who?”
According to Artificial Analysis, Muse Spark 1.3 scored 62 points on the Intelligence Index. The model ranked behind only Anthropic’s Claude Fable 5.1, which scored 66, and Claude Opus 5, which scored 63.
The Meta model also outperformed Google’s Gemini 1.5 Pro and Gemini 1.5 Flash, which scored 61 and 53, respectively.
Muse Spark 1.3 is Meta’s fourth Muse Spark model released in just five months. The latest model comes in two variants. Muse Spark 1.3 (max), currently in limited preview for Meta’s partners, scored 62. The publicly available Muse Spark 1.3 (xhigh) scored 61.
The xhigh variant tied with GPT-5.6 Sol (max), Grok 4.6 (high) and Claude Opus 5 (high). Its score stood four points above Muse Spark 1.2’s 57 in August and eight points above Muse Spark 1.1’s 53 in July.
Artificial Analysis said the latest Muse Spark variants recorded their biggest gains in agentic task performance and scientific reasoning.
Also Read: Meta Launches Muse Voice Transcribe with Multilingual AI Support
Muse Spark 1.3 (xhigh) rose from 35% to 47% on the Tau3-Bench Banking evaluation, while its Terminal-Bench 2.1 score increased from 80% to 85%. Its GDPval-AA v2 Elo rating also climbed from 1,615 to 1,709.
The max variant reached 52% on Tau3-Bench Banking, the highest score recorded by any model on that benchmark. It also achieved a GDPval-AA v2 Elo of 1,754.
On scientific reasoning, the xhigh variant’s CritPt score increased from 18% to 26%, while GPQA Diamond rose from 90% to 94%.
Muse Spark 1.3 also comes with a cost advantage. The xhigh variant costs about $0.55 per task, significantly less than GPT-5.6 Sol and Grok 4.6, which cost nearly double.
Artificial Analysis described Muse Spark as being on the “Pareto frontier” for intelligence versus cost, making it one of the most competitive models in terms of value.