Meta: Muse Spark 1.2

Meta: Muse Spark 1.2 is a paid multimodal model from Meta that accepts text, image, video, and file inputs. It offers a context window of 1,048,576 tokens and allows up to 943,718 completion tokens, so long documents, lengthy transcripts, and very large outputs can fit in a single request. It supports tool calling and reasoning. The listing does not specify structured output support or weight availability, so confirm those before building around them. Pricing is $1.25 per million input tokens and $4.25 per million output tokens, which makes it worth considering for long-context, video-capable workloads where per-token cost matters. Its blended benchmark score of 66.5 rests on a single independent benchmark (AI Index score of 65.3), which is thin evidence for ranking it against alternatives. Shortlist it for large-input multimodal tasks that use tools or reasoning, but run your own evaluations before committing.