Meta's Muse Spark 1.1 model has demonstrated superior performance in coding tasks compared to GLM-5.2, according to Artificial Analysis. The model scored 71.3 on the Coding Index, surpassing GLM-5.2's 68.8 and narrowly trailing GPT-5.6 Luna's 71.4. The improvement marks a significant leap for Meta's latest offering, which also shows enhanced cost efficiency.
Muse Spark 1.1 achieved its score after gaining eight points in just three months, primarily in coding and agent-based knowledge work. The model's cost per task is estimated at $0.26, compared to $0.37 for GLM-5.2 and $0.89 for GPT-5.4. Additionally, it uses only 94 million output tokens, less than GLM-5.2's 141 million. The hallucination rate has also dropped from 73 to 38 percent, with the model now more frequently declining to answer rather than providing incorrect information.
Meta has also expanded the context window to one million tokens, quadrupling the previous limit. The model is currently available only through Meta's own API. According to Artificial Analysis, Muse Spark 1.1 lands among the better value options with its score of 51 on the Intelligence Index and its cost-effective pricing.
Source: thedecoder
