Model comparison
MiMo-V2.5 vs Step 3
MiMo-V2.5 is the stronger model overall, scoring 43.4 to 40.5 on the Noometry Index.
Last verified . 16 shared benchmarks.
StepFunSummary
- They share 16 benchmarks with published results for both. MiMo-V2.5 scores higher in 8 categories and Step 3 in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where MiMo-V2.5 leads 61.6 to 54.3.
Side by side
| MiMo-V2.5 | Step 3 | |
|---|---|---|
| Provider | Xiaomi | StepFun |
| Noometry Index | 43.4 | 40.5 |
| Released | 2026-04-22 | — |
| Weights | Open | Open |
| Context window | 1.05M | — |
| Max output | 131K | — |
| Input $ / M tokens | $0.14 | — |
| Output $ / M tokens | $0.28 | — |
| Results tracked | 23 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.5 leads
MiMo-V2.5: 43.9 (#81), Step 3: 40.1 (#147)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Coding | 1469 | 1367 |
| LMArena WebDev | 1438 | — |
| SciCode | 43.1% | — |
| ALE-Bench | 513.95 | — |
Reasoning Too close to call
MiMo-V2.5: 28.6 (#101), Step 3: 28.4 (#105)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Hard Prompts | 1450 | 1355 |
| Kagi LLM Benchmark | — | 62.3% |
| CritPt | 3.7% | — |
Math Too close to call
MiMo-V2.5: 36.8 (#163), Step 3: 37.6 (#148)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Math | 1436 | 1366 |
| ProofBench | 16% | — |
Knowledge MiMo-V2.5 leads
MiMo-V2.5: 40.8 (#115), Step 3: 36.8 (#164)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Expert | 1460 | 1333 |
Multimodal MiMo-V2.5 leads
MiMo-V2.5: 39.8 (#54), Step 3: 35.5 (#86)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Vision | 1247 | 1177 |
Multilingual MiMo-V2.5 leads
MiMo-V2.5: 51.9 (#99), Step 3: 46.3 (#159)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Non-English | 1404 | 1327 |
| LMArena Chinese | 1468 | 1397 |
| LMArena German | 1421 | 1371 |
| LMArena Korean | 1363 | 1269 |
| LMArena Russian | 1395 | 1331 |
| LMArena Spanish | 1416 | 1371 |
| LMArena French | 1447 | — |
| LMArena Japanese | 1306 | — |
Instruction Following MiMo-V2.5 leads
MiMo-V2.5: 75.5 (#60), Step 3: 70.4 (#164)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1434 | 1332 |
Long Context MiMo-V2.5 leads
MiMo-V2.5: 44.2 (#73), Step 3: 40.3 (#157)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Longer Query | 1445 | 1326 |
Writing & Preference MiMo-V2.5 leads
MiMo-V2.5: 61.6 (#86), Step 3: 54.3 (#151)
| Benchmark | MiMo-V2.5 | Step 3 |
|---|---|---|
| LMArena Text | 1428 | 1350 |
| LMArena Creative Writing | 1393 | 1321 |
| LMArena Multi-Turn | 1445 | 1341 |
Frequently asked questions
Is MiMo-V2.5 better than Step 3?
MiMo-V2.5 is the stronger model overall, scoring 43.4 to 40.5 on the Noometry Index.
Is MiMo-V2.5 or Step 3 better for coding?
MiMo-V2.5 scores higher on coding benchmarks: 43.9 versus 40.1 in the Noometry coding category.
How many benchmarks do MiMo-V2.5 and Step 3 share?
16 benchmarks have published results for both models. MiMo-V2.5 has 23 scored results on Noometry and Step 3 has 17.