Model comparison
Muse Spark 1.3 vs Step 3.7 Flash
Muse Spark 1.3 is the stronger model overall, scoring 54.8 to 37.3 on the Noometry Index. Step 3.7 Flash costs 4.8× less per token, which makes it the better buy when Muse Spark 1.3's lead doesn't matter for your workload.
Last verified . 3 shared benchmarks.
Meta
StepFunSummary
- They share 3 benchmarks with published results for both. Muse Spark 1.3 scores higher in 3 categories and Step 3.7 Flash in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Muse Spark 1.3 leads 54.0 to 21.6.
- The biggest single-benchmark swing is NYT Connections (extended): 85.1% for Muse Spark 1.3 and 39.7% for Step 3.7 Flash.
- Step 3.7 Flash is cheaper at $0.18 / $1.11 per million input/output tokens, against $1.25 / $4.25 for Muse Spark 1.3.
- Muse Spark 1.3 accepts more context: 1.05M tokens versus 256K.
- Step 3.7 Flash has downloadable open weights; the other is API-only.
Side by side
| Muse Spark 1.3 | Step 3.7 Flash | |
|---|---|---|
| Provider | Meta | StepFun |
| Noometry Index | 54.8 | 37.3 |
| Released | 2026-09-02 | 2026-05-29 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 256K |
| Max output | 131K | 256K |
| Input $ / M tokens | $1.25 | $0.18 |
| Output $ / M tokens | $4.25 | $1.11 |
| Results tracked | 37 | 5 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Muse Spark 1.3 leads
Muse Spark 1.3: 56.6 (#21), Step 3.7 Flash: 40.0 (#150)
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| SciCode | 59.7% | 40% |
| CursorBench | 41.6% | — |
| LMArena WebDev | 1657 | — |
| LMArena Coding | 1514 | — |
| ALE-Bench | — | 694.12 |
Agentic & Tool Use Not comparable
Muse Spark 1.3: 38.6 (#30), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| APEX-Agents | 57.8% | — |
| GDP.pdf | 27.6% | — |
Reasoning Muse Spark 1.3 leads
Muse Spark 1.3: 54.0 (#27), Step 3.7 Flash: 21.6 (#219)
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| NYT Connections (extended) | 85.1% | 39.7% |
| CritPt | 26% | 2.3% |
| Chess Puzzles | 38% | — |
| LMArena Hard Prompts | 1503 | — |
| Mystery Game Puzzles | 25% | — |
| DTBench | 96.5% | — |
| LMCA | 53.9% | — |
| Bench to the Future 3 | 0.14 | — |
| Epoch Capabilities Index | 156.75 | — |
Math Muse Spark 1.3 leads
Muse Spark 1.3: 73.1 (#21), Step 3.7 Flash: 42.9 (#82)
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | 74.4% | — |
| FrontierMath Tier 4 | 46.3% | — |
| MathArena Final-Answer Competitions | — | 68.5% |
| OTIS Mock AIME 2024-2025 | 99.2% | — |
| ProofBench | 58% | — |
| LMArena Math | 1494 | — |
Knowledge Not comparable
Muse Spark 1.3: 42.6 (#95), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Expert | 1516 | — |
Multimodal Not comparable
Muse Spark 1.3: 43.7 (#22), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Vision | 1309 | — |
| LMArena Document | 1471 | — |
Multilingual Not comparable
Muse Spark 1.3: 57.4 (#8), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Non-English | 1481 | — |
| LMArena Chinese | 1529 | — |
| LMArena French | 1524 | — |
| LMArena German | 1515 | — |
| LMArena Japanese | 1474 | — |
| LMArena Korean | 1501 | — |
| LMArena Russian | 1490 | — |
| LMArena Spanish | 1490 | — |
Instruction Following Not comparable
Muse Spark 1.3: 77.5 (#22), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Instruction Following | 1477 | — |
Long Context Not comparable
Muse Spark 1.3: 45.6 (#32), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Longer Query | 1488 | — |
Writing & Preference Not comparable
Muse Spark 1.3: 73.6 (#9), Step 3.7 Flash: —
| Benchmark | Muse Spark 1.3 | Step 3.7 Flash |
|---|---|---|
| LMArena Text | 1490 | — |
| LMArena Creative Writing | 1455 | — |
| EQ-Bench Creative Writing | 1906 | — |
| LMArena Multi-Turn | 1482 | — |
Frequently asked questions
Is Muse Spark 1.3 better than Step 3.7 Flash?
Muse Spark 1.3 is the stronger model overall, scoring 54.8 to 37.3 on the Noometry Index. Step 3.7 Flash costs 4.8× less per token, which makes it the better buy when Muse Spark 1.3's lead doesn't matter for your workload.
Which is cheaper, Muse Spark 1.3 or Step 3.7 Flash?
Step 3.7 Flash is cheaper. It lists at $0.18 per million input tokens and $1.11 per million output tokens; Muse Spark 1.3 lists at $1.25 and $4.25.
Is Muse Spark 1.3 or Step 3.7 Flash better for coding?
Muse Spark 1.3 scores higher on coding benchmarks: 56.6 versus 40.0 in the Noometry coding category.
Which has the bigger context window?
Muse Spark 1.3 does, with 1.05M tokens against 256K.
How many benchmarks do Muse Spark 1.3 and Step 3.7 Flash share?
3 benchmarks have published results for both models. Muse Spark 1.3 has 37 scored results on Noometry and Step 3.7 Flash has 5.