Muse Ember
Muse Ember is Meta's smallest and most efficient reasoning model, built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long-document parsing. It pairs with Muse Wonder or Muse Spark, which plan while Muse Ember handles the volume.
Shogo Agent Index
UnrankedNot ranked yet: 1 agentic benchmark measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.
Agentic coding
68
estimated
Long-horizon autonomy
70
estimated
Knowledge-work agents
49
32% measured
Computer use
49
estimated
Security & context learning
57
estimated
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 84 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 89 | |
| DeepSWE | est. | 6% | predicted 67 | |
| FrontierSWE | est. | 5% | predicted 35 | |
| CursorBench | est. | 4% | predicted 46 | |
| GSO-Bench | est. | 4% | predicted 54 | |
| METR Time Horizons | est. | 10% | predicted 82 | |
| Vending-Bench 2 | est. | 6% | predicted 77 | |
| PostTrainBench | est. | 4% | predicted 32 | |
| APEX-Agents | 64.3% | 8% | source | |
| GDPval | est. | 7% | predicted 60 | |
| Remote Labor Index | est. | 5% | predicted 9 | |
| DeepResearch Bench | est. | 5% | predicted 51 | |
| OSWorld 2.0 | est. | 6% | predicted 23 | |
| OSWorld | est. | 4% | predicted 89 | |
| Cybench | est. | 5% | predicted 87 | |
| CL-bench | est. | 5% | predicted 26 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| SimpleBench | Reasoning | 72.7% | source |