Model comparison
Qwen3-30B-A3B vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 38.9 on the Noometry Index. Qwen3-30B-A3B costs 6.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 15 benchmarks with published results for both. Qwen3-30B-A3B scores higher in 1 category and Step 5 Preview in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 5 Preview leads 40.0 to 22.2.
- The biggest single-benchmark swing is SciCode: 33.3% for Qwen3-30B-A3B and 58.9% for Step 5 Preview.
- Qwen3-30B-A3B is cheaper at $0.12 / $0.50 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 41K.
- Qwen3-30B-A3B has downloadable open weights; the other is API-only.
Side by side
| Qwen3-30B-A3B | Step 5 Preview | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 38.9 | 47.9 |
| Released | 2025-04-28 | 2026-09-16 |
| Weights | Open | Proprietary |
| Context window | 41K | 1.02M |
| Max output | 16K | 66K |
| Input $ / M tokens | $0.12 | $1 |
| Output $ / M tokens | $0.50 | $2.70 |
| Results tracked | 32 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
Qwen3-30B-A3B: 37.5 (#194), Step 5 Preview: 51.8 (#36)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| SciCode | 33.3% | 58.9% |
| LMArena Coding | 1416 | 1480 |
| LMArena WebDev | — | 1564 |
| WeirdML | 29.8% | — |
Agentic & Tool Use Not comparable
Qwen3-30B-A3B: 29.8 (#82), Step 5 Preview: —
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| Berkeley Function Calling Leaderboard | 41.4% | — |
Reasoning Step 5 Preview leads
Qwen3-30B-A3B: 22.2 (#204), Step 5 Preview: 40.0 (#57)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| CritPt | 0.3% | 20.9% |
| LMArena Hard Prompts | 1398 | 1465 |
| Kagi LLM Benchmark | 54.9% | — |
| Chess Puzzles | 8% | — |
| DTBench | 69.3% | — |
| LMCA | 22.4% | — |
| Epoch Capabilities Index | 139.63 | — |
Math Step 5 Preview leads
Qwen3-30B-A3B: 37.4 (#157), Step 5 Preview: 46.1 (#72)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Math | 1394 | 1470 |
| MathArena Final-Answer Competitions | 47.8% | — |
| OTIS Mock AIME 2024-2025 | 70.3% | — |
| ProofBench | — | 42% |
Knowledge Too close to call
Qwen3-30B-A3B: 41.8 (#105), Step 5 Preview: 41.2 (#112)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1396 | 1470 |
| GPQA Diamond | 70.1% | — |
| Confabulations | 12.3% | — |
Multimodal Not comparable
Qwen3-30B-A3B: —, Step 5 Preview: 41.0 (#41)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Step 5 Preview leads
Qwen3-30B-A3B: 49.5 (#132), Step 5 Preview: 53.8 (#54)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1372 | 1432 |
| LMArena Chinese | 1433 | 1519 |
| LMArena Russian | 1370 | 1436 |
| LMArena Spanish | 1404 | 1439 |
| LMArena French | 1418 | — |
| LMArena German | 1380 | — |
| LMArena Japanese | 1337 | — |
| LMArena Korean | 1331 | — |
Instruction Following Step 5 Preview leads
Qwen3-30B-A3B: 72.0 (#142), Step 5 Preview: 76.0 (#50)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1363 | 1444 |
Long Context Step 5 Preview leads
Qwen3-30B-A3B: 31.0 (#283), Step 5 Preview: 44.5 (#64)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1379 | 1455 |
| Fiction.LiveBench | 40.6% | — |
Writing & Preference Step 5 Preview leads
Qwen3-30B-A3B: 55.6 (#143), Step 5 Preview: 62.8 (#71)
| Benchmark | Qwen3-30B-A3B | Step 5 Preview |
|---|---|---|
| LMArena Text | 1384 | 1442 |
| LMArena Creative Writing | 1317 | 1410 |
| LMArena Multi-Turn | 1378 | 1447 |
| Short-Story Creative Writing | 75.3% | — |
Frequently asked questions
Is Qwen3-30B-A3B better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 38.9 on the Noometry Index. Qwen3-30B-A3B costs 6.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, Qwen3-30B-A3B or Step 5 Preview?
Qwen3-30B-A3B is cheaper. It lists at $0.12 per million input tokens and $0.50 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is Qwen3-30B-A3B or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 37.5 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 41K.
How many benchmarks do Qwen3-30B-A3B and Step 5 Preview share?
15 benchmarks have published results for both models. Qwen3-30B-A3B has 32 scored results on Noometry and Step 5 Preview has 18.