Meta Muse Spark 1.1 matches GPT-5.5 on benchmarks, explainer says

Meta's Muse Spark 1.1 LLM scores on par with GPT-5.5 on reasoning, coding, and instruction-following benchmarks. The model is positioned as a hosted, high-performance alternative to OpenAI and Anthropic, accessible via Meta's tools or platforms like MindStudio.