Model comparison
Qwen3-Next 80B-A3B Instruct vs Step 3
Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 39.9 on the Noometry Index.
Last verified . 16 shared benchmarks.
Alibaba (Qwen)
StepFunSummary
- They share 16 benchmarks with published results for both. Qwen3-Next 80B-A3B Instruct scores higher in 7 categories and Step 3 in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Qwen3-Next 80B-A3B Instruct leads 52.3 to 46.5.
Side by side
| Qwen3-Next 80B-A3B Instruct | Step 3 | |
|---|---|---|
| Provider | Alibaba (Qwen) | StepFun |
| Noometry Index | 42.5 | 39.9 |
| Released | 2025-09 | — |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 33K | — |
| Input $ / M tokens | $0.50 | — |
| Output $ / M tokens | $2 | — |
| Results tracked | 25 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 41.9 (#101), Step 3: 39.5 (#144)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Coding | 1441 | 1367 |
Reasoning Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 30.1 (#79), Step 3: 27.6 (#102)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 66.7% | 62.3% |
| LMArena Hard Prompts | 1428 | 1355 |
Math Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 36.3 (#118), Step 3: 34.9 (#148)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Math | 1440 | 1366 |
| Omni-MATH | 46.7% | — |
Knowledge Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 41.3 (#102), Step 3: 36.2 (#161)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Expert | 1416 | 1333 |
| MMLU-Pro | 78.6% | — |
| Vectara Hallucination Rate | 9.3% | — |
| GPQA (HELM) | 63% | — |
Multimodal Not comparable
Qwen3-Next 80B-A3B Instruct: —, Step 3: 34.9 (#81)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Vision | — | 1176 |
Multilingual Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 52.3 (#88), Step 3: 46.5 (#154)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Non-English | 1407 | 1327 |
| LMArena Chinese | 1461 | 1396 |
| LMArena German | 1416 | 1371 |
| LMArena Korean | 1366 | 1271 |
| LMArena Russian | 1403 | 1331 |
| LMArena Spanish | 1436 | 1372 |
| LMArena French | 1415 | — |
| LMArena Japanese | 1398 | — |
Instruction Following Too close to call
Qwen3-Next 80B-A3B Instruct: 71.0 (#154), Step 3: 70.6 (#159)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1388 | 1332 |
| IFEval | 81% | — |
Long Context Step 3 leads
Qwen3-Next 80B-A3B Instruct: 37.0 (#218), Step 3: 40.3 (#152)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Longer Query | 1403 | 1326 |
| Fiction.LiveBench | 55.6% | — |
Writing & Preference Qwen3-Next 80B-A3B Instruct leads
Qwen3-Next 80B-A3B Instruct: 58.2 (#116), Step 3: 54.5 (#146)
| Benchmark | Qwen3-Next 80B-A3B Instruct | Step 3 |
|---|---|---|
| LMArena Text | 1417 | 1351 |
| LMArena Creative Writing | 1333 | 1322 |
| LMArena Multi-Turn | 1415 | 1340 |
| WildBench | 80.7% | — |
Frequently asked questions
Is Qwen3-Next 80B-A3B Instruct better than Step 3?
Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 39.9 on the Noometry Index.
Is Qwen3-Next 80B-A3B Instruct or Step 3 better for coding?
Qwen3-Next 80B-A3B Instruct scores higher on coding benchmarks: 41.9 versus 39.5 in the Noometry coding category.
How many benchmarks do Qwen3-Next 80B-A3B Instruct and Step 3 share?
16 benchmarks have published results for both models. Qwen3-Next 80B-A3B Instruct has 25 scored results on Noometry and Step 3 has 17.