Model comparison
Llama 13b vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 24.4 on the Noometry Index.
Last verified . 8 shared benchmarks.
Meta
StepFunSummary
- They share 8 benchmarks with published results for both. Llama 13b scores higher in 0 categories and Step 3 in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Step 3 leads 54.3 to 13.8.
Side by side
| Llama 13b | Step 3 | |
|---|---|---|
| Provider | Meta | StepFun |
| Noometry Index | 24.4 | 40.5 |
| Released | 2023-02-24 | — |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 21 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Llama 13b: 21.4 (#337), Step 3: 40.1 (#147)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Coding | 683 | 1367 |
Reasoning Step 3 leads
Llama 13b: 14.0 (#329), Step 3: 28.4 (#105)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Hard Prompts | 728 | 1355 |
| Kagi LLM Benchmark | — | 62.3% |
| BIG-Bench Hard | 37.9% | — |
| Epoch Capabilities Index | 100.58 | — |
| HellaSwag | 79.2% | — |
| LAMBADA | 75.2% | — |
| PIQA | 80.1% | — |
| WinoGrande | 73% | — |
Math Step 3 leads
Llama 13b: 26.7 (#256), Step 3: 37.6 (#148)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Math | 838 | 1366 |
| GSM8K | 20.6% | — |
Knowledge Not comparable
Llama 13b: —, Step 3: 36.8 (#164)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Expert | — | 1333 |
| ARC (AI2) Challenge | 52.7% | — |
| BoolQ | 78.7% | — |
| MMLU | 47.7% | — |
| OpenBookQA | 56.4% | — |
| TriviaQA | 77.9% | — |
Multimodal Not comparable
Llama 13b: —, Step 3: 35.5 (#86)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
| ScienceQA | 43.3% | — |
Multilingual Step 3 leads
Llama 13b: 16.6 (#297), Step 3: 46.3 (#159)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Non-English | 819 | 1327 |
| LMArena Chinese | — | 1397 |
| LMArena German | — | 1371 |
| LMArena Korean | — | 1269 |
| LMArena Russian | — | 1331 |
| LMArena Spanish | — | 1371 |
Instruction Following Step 3 leads
Llama 13b: 36.7 (#305), Step 3: 70.4 (#164)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Instruction Following | 781 | 1332 |
Long Context Not comparable
Llama 13b: —, Step 3: 40.3 (#157)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Longer Query | — | 1326 |
Writing & Preference Step 3 leads
Llama 13b: 13.8 (#312), Step 3: 54.3 (#151)
| Benchmark | Llama 13b | Step 3 |
|---|---|---|
| LMArena Text | 834 | 1350 |
| LMArena Creative Writing | 794 | 1321 |
| LMArena Multi-Turn | 753 | 1341 |
Frequently asked questions
Is Llama 13b better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 24.4 on the Noometry Index.
Is Llama 13b or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 21.4 in the Noometry coding category.
How many benchmarks do Llama 13b and Step 3 share?
8 benchmarks have published results for both models. Llama 13b has 21 scored results on Noometry and Step 3 has 17.