Muse Spark 1.3 vs Muse Wonder
Muse Wonder leads Muse Spark 1.3 by 6.2 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Muse Spark 1.3 wins 2 and Muse Wonder wins 10 of 12 shared benchmarks. Muse Spark 1.3 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Muse Spark 1.3 | Muse Wonder |
|---|---|---|
| Agentic coding | 60 | 68 |
| Long-horizon autonomy | 64 | 70 |
| Knowledge-work agents | 45 | 50 |
| Computer use | 43 | 50 |
| Security & context learning | 53 | 57 |
Shared benchmarks
| Benchmark | Muse Spark 1.3 | Muse Wonder |
|---|---|---|
| APEX-Agentsindex | 58.6% | 66.1% |
| CursorBenchindex | 41.6% | 50.4% |
| CritPt | 26.0% | 28.3% |
| DTBench | 96.5% | 95.2% |
| FrontierMath Tier 4 | 46.3% | 70.7% |
| FrontierMath Tiers 1-3 | 74.4% | 82.1% |
| LMCA | 53.9% | 56.8% |
| MirrorCode | 54.4% | 45.6% |
| OTIS Mock AIME | 99.2% | 99.4% |
| ProofBench | 58.0% | 84.0% |
| SciCode | 59.7% | 60.3% |
| WeirdML v3 | 6.7% | 15.1% |
Specs and price
| Muse Spark 1.3 | Muse Wonder | |
|---|---|---|
| Agent Index | 54.7 (#28) | 60.9 (#15) |
| Agent $/M | $0.030 | $1.69 |
| Input / output $/M | $0.10 / $0.20 | $4.00 / $20.00 |
| Context | 1M | 1M |
| Released | Sep 2, 2026 | Oct 12, 2026 |
| Weights | Closed | Closed |
| API hosts | 2 | 2 |
Don’t pick one. Route.
Shogo can run Muse Spark 1.3 and Muse Wonder side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo