Model comparison
Qwen3.8 27B vs Step 3.5 Flash
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index. Step 3.5 Flash costs 4.0× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 18 benchmarks with published results for both. Qwen3.8 27B scores higher in 7 categories and Step 3.5 Flash in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.8 27B leads 41.0 to 22.2.
- The biggest single-benchmark swing is NYT Connections (extended): 54.5% for Qwen3.8 27B and 28.4% for Step 3.5 Flash.
- Step 3.5 Flash is cheaper at $0.10 / $0.30 per million input/output tokens, against $0.04 / $2.30 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 256K.
Side by side
| Qwen3.8 27B | Step 3.5 Flash | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 46.0 | 42.3 |
| Released | 2026-08-14 | 2026-01-29 |
| Weights | Open | Open |
| Context window | 262K | 256K |
| Max output | 33K | 256K |
| Input $ / M tokens | $0.04 | $0.10 |
| Output $ / M tokens | $2.30 | $0.30 |
| Results tracked | 31 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Qwen3.8 27B: 50.5 (#44), Step 3.5 Flash: 42.4 (#105)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Coding | 1482 | 1436 |
| LMArena WebDev | 1593 | — |
| SciCode | 46.6% | — |
Agentic & Tool Use Not comparable
Qwen3.8 27B: 32.9 (#57), Step 3.5 Flash: —
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| APEX-Agents | 47.5% | — |
Reasoning Qwen3.8 27B leads
Qwen3.8 27B: 41.0 (#54), Step 3.5 Flash: 22.2 (#202)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| NYT Connections (extended) | 54.5% | 28.4% |
| LMArena Hard Prompts | 1460 | 1411 |
| ARC-AGI-2 | 42.4% | — |
| ARC-AGI-1 | 87.5% | — |
| CritPt | 5.4% | — |
| DTBench | 88% | — |
| LMCA | 41.4% | — |
| Surface Evolver Bench | 45% | — |
| Epoch Capabilities Index | 149.38 | — |
Math Step 3.5 Flash leads
Qwen3.8 27B: 37.1 (#161), Step 3.5 Flash: 42.6 (#84)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Math | 1456 | 1408 |
| MathArena Final-Answer Competitions | — | 66.8% |
| ProofBench | 16% | — |
Knowledge Qwen3.8 27B leads
Qwen3.8 27B: 41.6 (#109), Step 3.5 Flash: 39.6 (#132)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Expert | 1482 | 1421 |
Multimodal Not comparable
Qwen3.8 27B: 41.3 (#37), Step 3.5 Flash: —
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Vision | 1271 | — |
Multilingual Qwen3.8 27B leads
Qwen3.8 27B: 53.7 (#60), Step 3.5 Flash: 50.5 (#119)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Non-English | 1430 | 1385 |
| LMArena Chinese | 1504 | 1447 |
| LMArena French | 1465 | 1421 |
| LMArena German | 1438 | 1405 |
| LMArena Japanese | 1384 | 1354 |
| LMArena Korean | 1393 | 1352 |
| LMArena Russian | 1415 | 1385 |
| LMArena Spanish | 1448 | 1419 |
Instruction Following Qwen3.8 27B leads
Qwen3.8 27B: 75.8 (#53), Step 3.5 Flash: 73.1 (#124)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Instruction Following | 1439 | 1385 |
Long Context Qwen3.8 27B leads
Qwen3.8 27B: 44.3 (#70), Step 3.5 Flash: 42.8 (#117)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Longer Query | 1450 | 1402 |
Writing & Preference Qwen3.8 27B leads
Qwen3.8 27B: 65.8 (#43), Step 3.5 Flash: 58.8 (#113)
| Benchmark | Qwen3.8 27B | Step 3.5 Flash |
|---|---|---|
| LMArena Text | 1441 | 1403 |
| LMArena Creative Writing | 1384 | 1357 |
| LMArena Multi-Turn | 1441 | 1405 |
| EQ-Bench Creative Writing | 1671 | — |
Frequently asked questions
Is Qwen3.8 27B better than Step 3.5 Flash?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index. Step 3.5 Flash costs 4.0× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Which is cheaper, Qwen3.8 27B or Step 3.5 Flash?
Step 3.5 Flash is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Qwen3.8 27B lists at $0.04 and $2.30.
Is Qwen3.8 27B or Step 3.5 Flash better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 42.4 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 256K.
How many benchmarks do Qwen3.8 27B and Step 3.5 Flash share?
18 benchmarks have published results for both models. Qwen3.8 27B has 31 scored results on Noometry and Step 3.5 Flash has 19.