Muse Ember Hits OpenRouter for Agentic Coding

Muse Ember’s real pitch is economics: specialized subagents may matter more than another all-purpose frontier model.

• –Built for repetitive coding, tool calls, document analysis, and other structured agent steps
• –Offers a 1M-token context window and up to 512K output tokens, according to Meta
• –Meta reports 2.2-second p95 latency and 36–48% fewer output tokens in partner testing
• –Published Meta API pricing is $0.10 per million input tokens and $0.50 per million output tokens
• –Benchmark claims are largely internal or partner-reported, so independent testing will determine whether Ember can replace established coding subagents