Model comparison
Qwen3.7 Max vs Step 5 Preview
Qwen3.7 Max is the stronger model overall, scoring 51.5 to 47.9 on the Noometry Index. Step 5 Preview costs 2.6× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Last verified . 16 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 16 benchmarks with published results for both. Qwen3.7 Max scores higher in 7 categories and Step 5 Preview in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.7 Max leads 61.6 to 41.2.
- The biggest single-benchmark swing is ProofBench: 26% for Qwen3.7 Max and 42% for Step 5 Preview.
- Step 5 Preview is cheaper at $1 / $2.70 per million input/output tokens, against $2.50 / $7.50 for Qwen3.7 Max.
- Step 5 Preview accepts more context: 1.02M tokens versus 1M.
Side by side
| Qwen3.7 Max | Step 5 Preview | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 51.5 | 47.9 |
| Released | 2026-05-19 | 2026-09-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.02M |
| Max output | 131K | 66K |
| Input $ / M tokens | $2.50 | $1 |
| Output $ / M tokens | $7.50 | $2.70 |
| Results tracked | 33 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
Qwen3.7 Max: 50.4 (#45), Step 5 Preview: 51.8 (#36)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena WebDev | 1515 | 1564 |
| SciCode | 48.8% | 58.9% |
| LMArena Coding | 1498 | 1480 |
| SWE-bench Verified | 77.3% | — |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use Not comparable
Qwen3.7 Max: 22.1 (#135), Step 5 Preview: —
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| GBAEval | 0.4% | — |
Reasoning Qwen3.7 Max leads
Qwen3.7 Max: 49.2 (#38), Step 5 Preview: 40.0 (#57)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| CritPt | 13.4% | 20.9% |
| LMArena Hard Prompts | 1483 | 1465 |
| SimpleBench | 70.4% | — |
| NYT Connections (extended) | 85.1% | — |
| Chess Puzzles | 19% | — |
| EBR-Bench | 9.5% | — |
| Mystery Game Puzzles | 32% | — |
| DTBench | 92.3% | — |
| LMCA | 44% | — |
| Epoch Capabilities Index | 153.68 | — |
Math Qwen3.7 Max leads
Qwen3.7 Max: 62.4 (#32), Step 5 Preview: 46.1 (#72)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| ProofBench | 26% | 42% |
| LMArena Math | 1490 | 1470 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 34.1% | — |
| OTIS Mock AIME 2024-2025 | 95.6% | — |
Knowledge Qwen3.7 Max leads
Qwen3.7 Max: 61.6 (#28), Step 5 Preview: 41.2 (#112)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1488 | 1470 |
| GPQA Diamond | 90.9% | — |
| SimpleQA Verified | 55.8% | — |
Multimodal Not comparable
Qwen3.7 Max: —, Step 5 Preview: 41.0 (#41)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Qwen3.7 Max leads
Qwen3.7 Max: 56.9 (#15), Step 5 Preview: 53.8 (#54)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1474 | 1432 |
| LMArena Chinese | 1530 | 1519 |
| LMArena Russian | 1484 | 1436 |
| LMArena Spanish | — | 1439 |
Instruction Following Too close to call
Qwen3.7 Max: 76.7 (#38), Step 5 Preview: 76.0 (#50)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1460 | 1444 |
Long Context Too close to call
Qwen3.7 Max: 45.4 (#40), Step 5 Preview: 44.5 (#64)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1482 | 1455 |
Writing & Preference Qwen3.7 Max leads
Qwen3.7 Max: 65.0 (#54), Step 5 Preview: 62.8 (#71)
| Benchmark | Qwen3.7 Max | Step 5 Preview |
|---|---|---|
| LMArena Text | 1476 | 1442 |
| LMArena Creative Writing | 1449 | 1410 |
| LMArena Multi-Turn | 1481 | 1447 |
| EQ-Bench 4 | 1110 | — |
Frequently asked questions
Is Qwen3.7 Max better than Step 5 Preview?
Qwen3.7 Max is the stronger model overall, scoring 51.5 to 47.9 on the Noometry Index. Step 5 Preview costs 2.6× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Which is cheaper, Qwen3.7 Max or Step 5 Preview?
Step 5 Preview is cheaper. It lists at $1 per million input tokens and $2.70 per million output tokens; Qwen3.7 Max lists at $2.50 and $7.50.
Is Qwen3.7 Max or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 50.4 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 1M.
How many benchmarks do Qwen3.7 Max and Step 5 Preview share?
16 benchmarks have published results for both models. Qwen3.7 Max has 33 scored results on Noometry and Step 5 Preview has 18.