Meta released Muse Spark 1.1, an LLM matching GPT-5.5 on benchmarks

Independent benchmarks place Meta's Muse Spark 1.1 on par with GPT-5.5 across reasoning, coding, and instruction-following. Unlike the Llama series' open-weight focus, Muse Spark 1.1 is a hosted model optimized for complex reasoning and long-context tasks.