Model comparison
Qwen3 32B vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 39.2 on the Noometry Index.
Last verified . 14 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 14 benchmarks with published results for both. Qwen3 32B scores higher in 3 categories and Step 3 in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 3 leads 28.4 to 20.2.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 54.9% for Qwen3 32B and 62.3% for Step 3.
Side by side
| Qwen3 32B | Step 3 | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 39.2 | 40.5 |
| Released | 2025-04 | — |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 16K | — |
| Input $ / M tokens | $0.70 | — |
| Output $ / M tokens | $2.80 | — |
| Results tracked | 26 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Qwen3 32B: 37.7 (#190), Step 3: 40.1 (#147)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Coding | 1358 | 1367 |
| Aider Polyglot | 40% | — |
| SciCode | 35.4% | — |
Agentic & Tool Use Not comparable
Qwen3 32B: 32.6 (#62), Step 3: —
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 48.7% | — |
Reasoning Step 3 leads
Qwen3 32B: 20.2 (#241), Step 3: 28.4 (#105)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 54.9% | 62.3% |
| LMArena Hard Prompts | 1334 | 1355 |
| CritPt | 0.3% | — |
| Chess Puzzles | 5% | — |
| DTBench | 67.5% | — |
| LMCA | 17.3% | — |
| Epoch Capabilities Index | 138.51 | — |
Math Qwen3 32B leads
Qwen3 32B: 39.7 (#99), Step 3: 37.6 (#148)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Math | 1399 | 1366 |
| OTIS Mock AIME 2024-2025 | 66.9% | — |
Knowledge Qwen3 32B leads
Qwen3 32B: 40.0 (#125), Step 3: 36.8 (#164)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Expert | 1362 | 1333 |
| GPQA Diamond | 65.7% | — |
| Vectara Hallucination Rate | 5.9% | — |
Multimodal Not comparable
Qwen3 32B: —, Step 3: 35.5 (#86)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Too close to call
Qwen3 32B: 45.6 (#167), Step 3: 46.3 (#159)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Non-English | 1317 | 1327 |
| LMArena Chinese | 1357 | 1397 |
| LMArena German | 1341 | 1371 |
| LMArena Russian | 1311 | 1331 |
| LMArena Korean | — | 1269 |
| LMArena Spanish | — | 1371 |
Instruction Following Step 3 leads
Qwen3 32B: 68.9 (#179), Step 3: 70.4 (#164)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1305 | 1332 |
Long Context Qwen3 32B leads
Qwen3 32B: 43.8 (#87), Step 3: 40.3 (#157)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Longer Query | 1327 | 1326 |
| Fiction.LiveBench | 74.2% | — |
Writing & Preference Step 3 leads
Qwen3 32B: 52.9 (#163), Step 3: 54.3 (#151)
| Benchmark | Qwen3 32B | Step 3 |
|---|---|---|
| LMArena Text | 1340 | 1350 |
| LMArena Creative Writing | 1297 | 1321 |
| LMArena Multi-Turn | 1331 | 1341 |
Frequently asked questions
Is Qwen3 32B better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 39.2 on the Noometry Index.
Is Qwen3 32B or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 37.7 in the Noometry coding category.
How many benchmarks do Qwen3 32B and Step 3 share?
14 benchmarks have published results for both models. Qwen3 32B has 26 scored results on Noometry and Step 3 has 17.