Meta launches Muse Ember for low-cost AI agent tasks

Meta launched Muse Ember on November 9, presenting it as a low-cost model for the repetitive tasks handled by artificial intelligence agents, including tool calls, routing, information extraction, document processing and smaller coding jobs. The company said the model is available through the Meta Model API and its Muse Code product.

Meta lists Ember at $0.10 per million input tokens and $0.50 per million output tokens. The model can process text, images, audio and video, with a context window of up to one million tokens, according to the company.

Meta’s cost and capability comparisons rely in part on Artificial Analysis benchmarks. On the firm’s Intelligence Index, Ember and Anthropic’s Claude Haiku 5.5 each scored 36. Meta-reported estimates put the cost of the benchmark task at $0.14 for Ember and $0.41 for Haiku. Those figures are estimated costs for the benchmark, not fixed customer charges.

Results varied by test. On FrontierCode 2.0, which evaluates code changes, Haiku scored 34.9 while Ember scored 27.2. The launch reports also cited higher Ember scores on other agent and coding benchmarks, though some results came from Meta’s own tests. The available comparisons do not...