Meta launches Muse Ember for cost-efficient AI agents
Meta introduced Muse Ember, an entry-level AI model priced at 10 cents per million input tokens and 50 cents per million output tokens, with cached prompts dropping to 0.4 cents. It is the first model Meta trained using its flagship Muse Wonder LLM. Benchmarked against Anthropic's Claude Haiku 5.5 and OpenAI's GPT-6 Luna, Muse Ember matched Haiku's score at roughly a third the cost and outperformed both rivals on seven of nine benchmarks, including a 17% lead on business automation tasks. The model supports a 1 million token context window and multimodal inputs.
