Meta AI Releases Muse Spark 1.2 (xHigh) with High Cost-Performance Efficiency
Meta AI has released Muse Spark 1.2 (xHigh), which ranks fourth in the Text Arena benchmark with 1,498 points and eleventh in the Vision Arena with 1,290 points. The model is priced competitively at $1.25 per million input tokens and $4.25 per million output tokens. By achieving high benchmark scores at a low price point, Muse Spark 1.2 (xHigh) establishes a new Pareto frontier for cost-performance, making advanced LLM capabilities more accessible for developers. It also shows significant improvements in specialized categories like HTML, gaming, and frontend development compared to version 1.1. Compared to Muse Spark 1.1, version 1.2 saw its biggest ranking gains in HTML (moving from #24 to #8) and Gaming (moving from #23 to #13). The model is developed by Meta Superintelligence Labs (MSL) as part of their focus on multimodal reasoning and agentic tasks.
## BACKGROUND
Muse Spark is a large language model family developed by Meta, designed for multimodal reasoning, coding, and AI-assisted tasks. The Text Arena is an interactive benchmark that evaluates LLMs on reasoning, coding, and general tasks, while the Pareto frontier in machine learning represents the optimal trade-offs between competing objectives, such as model performance and operational cost.