Model comparison
Qwen3.5 27B vs Step 3.7 Flash
Qwen3.5 27B is the stronger model overall, scoring 41.9 to 37.3 on the Noometry Index. Step 3.7 Flash costs 2.0× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Last verified . 3 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 3 benchmarks with published results for both. Qwen3.5 27B scores higher in 1 category and Step 3.7 Flash in 2 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.5 27B leads 27.5 to 21.6.
- The biggest single-benchmark swing is MathArena Final-Answer Competitions: 56.7% for Qwen3.5 27B and 68.5% for Step 3.7 Flash.
- Step 3.7 Flash is cheaper at $0.18 / $1.11 per million input/output tokens, against $0.30 / $2.40 for Qwen3.5 27B.
- Qwen3.5 27B accepts more context: 262K tokens versus 256K.
Side by side
| Qwen3.5 27B | Step 3.7 Flash | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 41.9 | 37.3 |
| Released | 2026-02-23 | 2026-05-29 |
| Weights | Open | Open |
| Context window | 262K | 256K |
| Max output | 66K | 256K |
| Input $ / M tokens | $0.30 | $0.18 |
| Output $ / M tokens | $2.40 | $1.11 |
| Results tracked | 28 | 5 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3.7 Flash leads
Qwen3.5 27B: 38.9 (#168), Step 3.7 Flash: 40.0 (#150)
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| ALE-Bench | 349.45 | 694.12 |
| LMArena WebDev | 1358 | — |
| SciCode | — | 40% |
| WeirdML | 39.5% | — |
| LMArena Coding | 1427 | — |
Agentic & Tool Use Not comparable
Qwen3.5 27B: —, Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| Vending-Bench 2 | 201.98 | — |
Reasoning Qwen3.5 27B leads
Qwen3.5 27B: 27.5 (#117), Step 3.7 Flash: 21.6 (#219)
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| NYT Connections (extended) | 47.9% | 39.7% |
| CritPt | — | 2.3% |
| Thematic Generalization | 45.5% | — |
| LMArena Hard Prompts | 1414 | — |
| DTBench | 82.4% | — |
| LMCA | 34% | — |
Math Step 3.7 Flash leads
Qwen3.5 27B: 38.8 (#127), Step 3.7 Flash: 42.9 (#82)
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| MathArena Final-Answer Competitions | 56.7% | 68.5% |
| LMArena Math | 1429 | — |
Knowledge Not comparable
Qwen3.5 27B: 38.0 (#150), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| Vectara Hallucination Rate | 12.1% | — |
| LMArena Expert | 1428 | — |
Multimodal Not comparable
Qwen3.5 27B: 39.4 (#59), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| LMArena Vision | 1241 | — |
Multilingual Not comparable
Qwen3.5 27B: 50.8 (#115), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| LMArena Non-English | 1390 | — |
| LMArena Chinese | 1478 | — |
| LMArena French | 1410 | — |
| LMArena German | 1393 | — |
| LMArena Japanese | 1345 | — |
| LMArena Korean | 1358 | — |
| LMArena Russian | 1390 | — |
| LMArena Spanish | 1407 | — |
Instruction Following Not comparable
Qwen3.5 27B: 73.5 (#119), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| LMArena Instruction Following | 1393 | — |
Long Context Not comparable
Qwen3.5 27B: 43.1 (#106), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| LMArena Longer Query | 1413 | — |
Writing & Preference Not comparable
Qwen3.5 27B: 59.3 (#111), Step 3.7 Flash: —
| Benchmark | Qwen3.5 27B | Step 3.7 Flash |
|---|---|---|
| LMArena Text | 1409 | — |
| LMArena Creative Writing | 1362 | — |
| LMArena Multi-Turn | 1410 | — |
Frequently asked questions
Is Qwen3.5 27B better than Step 3.7 Flash?
Qwen3.5 27B is the stronger model overall, scoring 41.9 to 37.3 on the Noometry Index. Step 3.7 Flash costs 2.0× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Which is cheaper, Qwen3.5 27B 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; Qwen3.5 27B lists at $0.30 and $2.40.
Is Qwen3.5 27B or Step 3.7 Flash better for coding?
Step 3.7 Flash scores higher on coding benchmarks: 40.0 versus 38.9 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5 27B does, with 262K tokens against 256K.
How many benchmarks do Qwen3.5 27B and Step 3.7 Flash share?
3 benchmarks have published results for both models. Qwen3.5 27B has 28 scored results on Noometry and Step 3.7 Flash has 5.