research.meta.aiIntroducing Our Smallest, Most Efficient Model: Muse Ember@eddie · nov 10Listen10:00Save1 saveMeta released Muse Ember, its smallest and most efficient model, beating Claude Haiku 5.5 and GPT-6 Luna on intelligence-per-cost on agentic coding and reasoning tasks at roughly $0.14 per task.More onMeta172GPT-6 Luna5Muse Wonder3Artificial Analysis2Muse Spark2Claude Haiku 5.51#developer-tools4938#ai-coding2039#ai-agents1480#ai-models4#cost-efficiency1RelatedMeta's AI Chief On AI Beef, New Models And Life With Zuck - EP 71 Alex WangMeta hired Scale AI co-founder Alex Wang to lead its AI resurgence, resulting in the Muse Spark model as Meta attempts to catch OpenAI and Anthropic.also on Muse Spark, MetaQwenQwen4-Max scores 43 on Artificial Analysis, showing improved coding and agentic workflows over Qwen 3.8 but remains behind proprietary frontier models like Opus.also on Artificial Analysis, #ai-codingMuse Spark OpenMuse Spark Open releases the Muse Spark creative toolkit as open source, inviting developers to extend and integrate its features into their own applications.also on Muse Spark, #developer-toolsMuse Wonder: Zuckerberg and Wang's $100B AIZuckerberg and Wang reveal Meta’s $100B+ AI investment, detail Muse Wonder’s active memory, multimodal perception, and frontier coding, and outline their vision for personal AI agents.also on Muse Wonder, Meta, #ai-agents, #ai-coding, #developer-toolsMeta Unveils New Series of Advanced A.I. ModelsMeta announced two new advanced AI models, Muse Wonder and Muse Wonder Full, designed to compete with leading models from Anthropic and OpenAI.also on Muse Wonder, Meta