Arena.ai's code model evaluation platform ranking models on WebDev and other tasks by performance and cost.
Meta's Muse Ember (Max) tops the Code Arena WebDev small-model leaderboard at 1608 points, beating Claude Haiku 5.5 and GPT-6 Luna. At identical API pricing, its median cost per task is just $0.03, roughly half of Luna and a quarter of Haiku, placing it on the cost-performance Pareto frontier. Generations averaged about two minutes each, producing 20+ websites and animations in a single pass.
Meta has introduced Muse Ember, its smallest and most efficient model to date, targeting complex reasoning execution and everyday agent workflows with cross-app operation, sequential tool calling, test running and targeted code fixes, plus a 1M-token context and native multimodal understanding. It tops the Code Arena WebDev small-model ranking at 1608 points, beating Claude Haiku 5.5 and GPT-6 Luna, while sitting on the Pareto frontier at a median cost of just $0.03 per task.