Model comparison
Llama 3.1-405B vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 30.7 on the Noometry Index.
Last verified . 16 shared benchmarks.
Meta
StepFunSummary
- They share 16 benchmarks with published results for both. Llama 3.1-405B scores higher in 0 categories and Step 3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Step 3 leads 37.6 to 18.4.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 45% for Llama 3.1-405B and 62.3% for Step 3.
Side by side
| Llama 3.1-405B | Step 3 | |
|---|---|---|
| Provider | Meta | StepFun |
| Noometry Index | 30.7 | 40.5 |
| Released | 2024-07-23 | — |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 42 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Llama 3.1-405B: 33.1 (#262), Step 3: 40.1 (#147)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Coding | 1291 | 1367 |
| WeirdML | 21.4% | — |
Agentic & Tool Use Not comparable
Llama 3.1-405B: 21.0 (#140), Step 3: —
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| TheAgentCompany | 7.4% | — |
| Cybench | 7.5% | — |
Reasoning Step 3 leads
Llama 3.1-405B: 16.8 (#300), Step 3: 28.4 (#105)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 45% | 62.3% |
| LMArena Hard Prompts | 1269 | 1355 |
| SimpleBench | 23% | — |
| DTBench | 61.4% | — |
| BIG-Bench Hard | 82.9% | — |
| Epoch Capabilities Index | 128.75 | — |
| ForecastBench | 59.9 | — |
| HellaSwag | 89.2% | — |
| PIQA | 85.9% | — |
| WinoGrande | 89.2% | — |
Math Step 3 leads
Llama 3.1-405B: 18.4 (#290), Step 3: 37.6 (#148)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Math | 1281 | 1366 |
| OTIS Mock AIME 2024-2025 | 9.7% | — |
| Omni-MATH | 24.9% | — |
| MATH Level 5 | 49.8% | — |
Knowledge Step 3 leads
Llama 3.1-405B: 30.4 (#227), Step 3: 36.8 (#164)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Expert | 1243 | 1333 |
| GPQA Diamond | 50.9% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 17.6% | — |
| GPQA (HELM) | 52.2% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 84.5% | — |
| TriviaQA | 82.7% | — |
Multimodal Not comparable
Llama 3.1-405B: —, Step 3: 35.5 (#86)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Step 3 leads
Llama 3.1-405B: 40.7 (#214), Step 3: 46.3 (#159)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Non-English | 1248 | 1327 |
| LMArena Chinese | 1242 | 1397 |
| LMArena German | 1252 | 1371 |
| LMArena Korean | 1184 | 1269 |
| LMArena Russian | 1265 | 1331 |
| LMArena Spanish | 1260 | 1371 |
| LMArena French | 1279 | — |
| LMArena Japanese | 1208 | — |
Instruction Following Step 3 leads
Llama 3.1-405B: 65.9 (#214), Step 3: 70.4 (#164)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1259 | 1332 |
| IFEval | 81.1% | — |
Long Context Step 3 leads
Llama 3.1-405B: 38.4 (#197), Step 3: 40.3 (#157)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Longer Query | 1266 | 1326 |
Writing & Preference Step 3 leads
Llama 3.1-405B: 38.9 (#251), Step 3: 54.3 (#151)
| Benchmark | Llama 3.1-405B | Step 3 |
|---|---|---|
| LMArena Text | 1284 | 1350 |
| LMArena Creative Writing | 1262 | 1321 |
| LMArena Multi-Turn | 1297 | 1341 |
| EQ-Bench Creative Writing | 870 | — |
| WildBench | 78.3% | — |
Frequently asked questions
Is Llama 3.1-405B better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 30.7 on the Noometry Index.
Is Llama 3.1-405B or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 33.1 in the Noometry coding category.
How many benchmarks do Llama 3.1-405B and Step 3 share?
16 benchmarks have published results for both models. Llama 3.1-405B has 42 scored results on Noometry and Step 3 has 17.