Meta launched Muse Ember, the smallest and lowest-priced model in its Muse lineup, built for the routine work AI agents repeat, such as calling tools, fixing small pieces of code, sorting requests and pulling data out of files. Ember costs $0.10 per million input tokens and $0.50 per million output tokens, and Artificial Analysis puts its cost per task at about $0.14, compared with $2.39 for Meta’s Muse Spark 1.3, which scored one point higher. Meta says Ember beat Claude Haiku 5.5 and GPT-6 Luna on business automation and computer-use benchmarks, but Haiku still leads on real-world software engineering tests. The model takes images, audio and video as input, has a 1M-token context window, and is available through the Meta Model API, Muse Code, Google Cloud and Oracle Cloud.
