Model comparison
Gemma 2B vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 29.6 on the Noometry Index.
Last verified . 11 shared benchmarks.
Google
StepFunSummary
- They share 11 benchmarks with published results for both. Gemma 2B scores higher in 0 categories and Step 3 in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Step 3 leads 54.3 to 24.0.
Side by side
| Gemma 2B | Step 3 | |
|---|---|---|
| Provider | StepFun | |
| Noometry Index | 29.6 | 40.5 |
| Released | 2024-02-21 | — |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 23 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Gemma 2B: 29.4 (#305), Step 3: 40.1 (#147)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Coding | 1010 | 1367 |
| HumanEval+ | 20.7% | — |
| MBPP+ | 34.1% | — |
Reasoning Step 3 leads
Gemma 2B: 18.8 (#275), Step 3: 28.4 (#105)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Hard Prompts | 989 | 1355 |
| Kagi LLM Benchmark | — | 62.3% |
| BIG-Bench Hard | 35.2% | — |
| Epoch Capabilities Index | 94.2 | — |
| HellaSwag | 71.4% | — |
| PIQA | 77.3% | — |
| WinoGrande | 65.4% | — |
Math Step 3 leads
Gemma 2B: 30.0 (#239), Step 3: 37.6 (#148)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Math | 1009 | 1366 |
| GSM8K | 17.7% | — |
Knowledge Not comparable
Gemma 2B: —, Step 3: 36.8 (#164)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Expert | — | 1333 |
| ARC (AI2) Challenge | 42.1% | — |
| BoolQ | 69.4% | — |
| MMLU | 42.3% | — |
| TriviaQA | 53.2% | — |
Multimodal Not comparable
Gemma 2B: —, Step 3: 35.5 (#86)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Step 3 leads
Gemma 2B: 23.0 (#294), Step 3: 46.3 (#159)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Non-English | 958 | 1327 |
| LMArena Chinese | 986 | 1397 |
| LMArena Russian | 937 | 1331 |
| LMArena German | — | 1371 |
| LMArena Korean | — | 1269 |
| LMArena Spanish | — | 1371 |
Instruction Following Step 3 leads
Gemma 2B: 48.5 (#302), Step 3: 70.4 (#164)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Instruction Following | 970 | 1332 |
Long Context Step 3 leads
Gemma 2B: 29.9 (#291), Step 3: 40.3 (#157)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Longer Query | 981 | 1326 |
Writing & Preference Step 3 leads
Gemma 2B: 24.0 (#308), Step 3: 54.3 (#151)
| Benchmark | Gemma 2B | Step 3 |
|---|---|---|
| LMArena Text | 1002 | 1350 |
| LMArena Creative Writing | 987 | 1321 |
| LMArena Multi-Turn | 945 | 1341 |
Frequently asked questions
Is Gemma 2B better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 29.6 on the Noometry Index.
Is Gemma 2B or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 29.4 in the Noometry coding category.
How many benchmarks do Gemma 2B and Step 3 share?
11 benchmarks have published results for both models. Gemma 2B has 23 scored results on Noometry and Step 3 has 17.