Model comparison
Gemma 3 4B vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 28.1 on the Noometry Index.
Last verified . 13 shared benchmarks.
Google
StepFunSummary
- They share 13 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and Step 3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Step 3 leads 36.8 to 11.8.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 25.2% for Gemma 3 4B and 62.3% for Step 3.
Side by side
| Gemma 3 4B | Step 3 | |
|---|---|---|
| Provider | StepFun | |
| Noometry Index | 28.1 | 40.5 |
| Released | 2025-03-12 | — |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.04 | — |
| Output $ / M tokens | $0.08 | — |
| Results tracked | 22 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Gemma 3 4B: 35.9 (#215), Step 3: 40.1 (#147)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Coding | 1230 | 1367 |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.9 (#142), Step 3: —
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
Reasoning Step 3 leads
Gemma 3 4B: 13.2 (#335), Step 3: 28.4 (#105)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 25.2% | 62.3% |
| LMArena Hard Prompts | 1253 | 1355 |
| Chess Puzzles | 0% | — |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| Epoch Capabilities Index | 116.02 | — |
Math Step 3 leads
Gemma 3 4B: 16.8 (#292), Step 3: 37.6 (#148)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Math | 1239 | 1366 |
| OTIS Mock AIME 2024-2025 | 7.5% | — |
Knowledge Step 3 leads
Gemma 3 4B: 11.8 (#299), Step 3: 36.8 (#164)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Expert | 1223 | 1333 |
| GPQA Diamond | 23.2% | — |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, Step 3: 35.5 (#86)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Step 3 leads
Gemma 3 4B: 42.5 (#194), Step 3: 46.3 (#159)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Non-English | 1273 | 1327 |
| LMArena German | 1281 | 1371 |
| LMArena Russian | 1294 | 1331 |
| LMArena Chinese | — | 1397 |
| LMArena Korean | — | 1269 |
| LMArena Spanish | — | 1371 |
Instruction Following Step 3 leads
Gemma 3 4B: 65.2 (#225), Step 3: 70.4 (#164)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1332 |
Long Context Step 3 leads
Gemma 3 4B: 38.7 (#194), Step 3: 40.3 (#157)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Longer Query | 1273 | 1326 |
Writing & Preference Step 3 leads
Gemma 3 4B: 42.0 (#239), Step 3: 54.3 (#151)
| Benchmark | Gemma 3 4B | Step 3 |
|---|---|---|
| LMArena Text | 1291 | 1350 |
| LMArena Creative Writing | 1271 | 1321 |
| LMArena Multi-Turn | 1255 | 1341 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 28.1 on the Noometry Index.
Is Gemma 3 4B or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 35.9 in the Noometry coding category.
How many benchmarks do Gemma 3 4B and Step 3 share?
13 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Step 3 has 17.