Model comparison
Qwen3.6 35B-A3B vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 37.6 on the Noometry Index. Qwen3.6 35B-A3B costs 2.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 2 benchmarks with published results for both. Qwen3.6 35B-A3B scores higher in 1 category and Step 5 Preview in 3 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in coding, where Step 5 Preview leads 51.8 to 37.2.
- The biggest single-benchmark swing is SciCode: 35.8% for Qwen3.6 35B-A3B and 58.9% for Step 5 Preview.
- Qwen3.6 35B-A3B is cheaper at $0.25 / $1.49 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 262K.
- Qwen3.6 35B-A3B has downloadable open weights; the other is API-only.
Side by side
| Qwen3.6 35B-A3B | Step 5 Preview | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 37.6 | 47.9 |
| Released | 2026-04-01 | 2026-09-16 |
| Weights | Open | Proprietary |
| Context window | 262K | 1.02M |
| Max output | 66K | 66K |
| Input $ / M tokens | $0.25 | $1 |
| Output $ / M tokens | $1.49 | $2.70 |
| Results tracked | 14 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
Qwen3.6 35B-A3B: 37.2 (#196), Step 5 Preview: 51.8 (#36)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| SciCode | 35.8% | 58.9% |
| LMArena WebDev | — | 1564 |
| WeirdML | 34.5% | — |
| LMArena Coding | — | 1480 |
Agentic & Tool Use Not comparable
Qwen3.6 35B-A3B: 22.1 (#134), Step 5 Preview: —
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| Terminal-Bench | 23% | — |
Reasoning Step 5 Preview leads
Qwen3.6 35B-A3B: 28.0 (#109), Step 5 Preview: 40.0 (#57)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| CritPt | 0.3% | 20.9% |
| NYT Connections (extended) | 41.6% | — |
| Chess Puzzles | 26% | — |
| LMArena Hard Prompts | — | 1465 |
| Mystery Game Puzzles | 22% | — |
| DTBench | 73.9% | — |
| LMCA | 29.7% | — |
| Surface Evolver Bench | 44.4% | — |
| Epoch Capabilities Index | 143.93 | — |
Math Step 5 Preview leads
Qwen3.6 35B-A3B: 38.9 (#121), Step 5 Preview: 46.1 (#72)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| FrontierMath (Tiers 1-3) | 20.4% | — |
| OTIS Mock AIME 2024-2025 | 86.7% | — |
| ProofBench | — | 42% |
| LMArena Math | — | 1470 |
Knowledge Qwen3.6 35B-A3B leads
Qwen3.6 35B-A3B: 51.3 (#68), Step 5 Preview: 41.2 (#112)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| GPQA Diamond | 84.8% | — |
| LMArena Expert | — | 1470 |
Multimodal Not comparable
Qwen3.6 35B-A3B: —, Step 5 Preview: 41.0 (#41)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Not comparable
Qwen3.6 35B-A3B: —, Step 5 Preview: 53.8 (#54)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Non-English | — | 1432 |
| LMArena Chinese | — | 1519 |
| LMArena Russian | — | 1436 |
| LMArena Spanish | — | 1439 |
Instruction Following Not comparable
Qwen3.6 35B-A3B: —, Step 5 Preview: 76.0 (#50)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | — | 1444 |
Long Context Not comparable
Qwen3.6 35B-A3B: —, Step 5 Preview: 44.5 (#64)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | — | 1455 |
Writing & Preference Not comparable
Qwen3.6 35B-A3B: —, Step 5 Preview: 62.8 (#71)
| Benchmark | Qwen3.6 35B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Text | — | 1442 |
| LMArena Creative Writing | — | 1410 |
| LMArena Multi-Turn | — | 1447 |
Frequently asked questions
Is Qwen3.6 35B-A3B better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 37.6 on the Noometry Index. Qwen3.6 35B-A3B costs 2.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, Qwen3.6 35B-A3B or Step 5 Preview?
Qwen3.6 35B-A3B is cheaper. It lists at $0.25 per million input tokens and $1.49 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is Qwen3.6 35B-A3B or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 37.2 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 262K.
How many benchmarks do Qwen3.6 35B-A3B and Step 5 Preview share?
2 benchmarks have published results for both models. Qwen3.6 35B-A3B has 14 scored results on Noometry and Step 5 Preview has 18.