Muse Ember (xHigh): benchmark results
Provider: Meta. Released 2026-11-09. Access: API.
Aggregate Rating 1816 ± 16, rank #117 of 2201 rated models, from 66 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA-Omniscience v1.1 Hallucination Rate | 17.34 | Hallucination Rate (%) | 97 |
| AA GDP.pdf - Real Estate - Criterion Pass Rate | 76.42 | Criterion Pass Rate (%) | 94.1 |
| AA-Briefcase - PDF - Rubric Pass Rate | 62.07 | Rubric Pass Rate (%) | 93.8 |
| AutomationBench-AA - HR - Raw Objective Completion | 76.98 | Raw Objective Completion (%) | 92.2 |
| AutomationBench-AA - Sales - Raw Objective Completion | 87.36 | Raw Objective Completion (%) | 92.2 |
| AutomationBench-AA - Strict | 37.44 | Strict Task Success (%) | 92 |
| AutomationBench-AA - Marketing - Raw Objective Completion | 88.55 | Raw Objective Completion (%) | 91.8 |
| AutomationBench-AA - Operations - Raw Objective Completion | 96.02 | Raw Objective Completion (%) | 91.6 |
| AutomationBench-AA | 66.71 | Guardrail-Adjusted Objective Completion (%) | 91.4 |
| AutomationBench-AA - Support - Raw Objective Completion | 90.58 | Raw Objective Completion (%) | 91.4 |
| AA-Briefcase - DOCX - Rubric Pass Rate | 56.8 | Rubric Pass Rate (%) | 91.2 |
| AutomationBench-AA - Finance - Raw Objective Completion | 88.79 | Raw Objective Completion (%) | 90.6 |
Interactive version: theaggregate.ai/model?slug=muse-ember-xhigh · How It Works · Data refreshed daily, snapshot 2026-11-12.