Model comparison
Nova Premier 1.0 vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 38.3 on the Noometry Index.
Last verified . 1 shared benchmarks.
Summary
- They share 1 benchmark with published results for both. Nova Premier 1.0 scores higher in 0 categories and Step 3 in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 3 leads 28.4 to 23.0.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 44.8% for Nova Premier 1.0 and 62.3% for Step 3.
- Step 3 has downloadable open weights; the other is API-only.
Side by side
| Nova Premier 1.0 | Step 3 | |
|---|---|---|
| Provider | Amazon | StepFun |
| Noometry Index | 38.3 | 40.5 |
| Released | 2025-04-30 | — |
| Weights | Proprietary | Open |
| Context window | 1M | — |
| Max output | 10K | — |
| Input $ / M tokens | $2.50 | — |
| Output $ / M tokens | $12.50 | — |
| Results tracked | 6 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Nova Premier 1.0: —, Step 3: 40.1 (#147)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| LMArena Coding | — | 1367 |
Reasoning Step 3 leads
Nova Premier 1.0: 23.0 (#185), Step 3: 28.4 (#105)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| Kagi LLM Benchmark | 44.8% | 62.3% |
| LMArena Hard Prompts | — | 1355 |
Math Step 3 leads
Nova Premier 1.0: 33.8 (#199), Step 3: 37.6 (#148)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| Omni-MATH | 35% | — |
| LMArena Math | — | 1366 |
Knowledge Step 3 leads
Nova Premier 1.0: 35.8 (#180), Step 3: 36.8 (#164)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| MMLU-Pro | 72.6% | — |
| GPQA (HELM) | 51.8% | — |
| LMArena Expert | — | 1333 |
Multimodal Not comparable
Nova Premier 1.0: —, Step 3: 35.5 (#86)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Not comparable
Nova Premier 1.0: —, Step 3: 46.3 (#159)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| LMArena Non-English | — | 1327 |
| LMArena Chinese | — | 1397 |
| LMArena German | — | 1371 |
| LMArena Korean | — | 1269 |
| LMArena Russian | — | 1331 |
| LMArena Spanish | — | 1371 |
Instruction Following Step 3 leads
Nova Premier 1.0: 66.8 (#204), Step 3: 70.4 (#164)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| IFEval | 80.3% | — |
| LMArena Instruction Following | — | 1332 |
Long Context Not comparable
Nova Premier 1.0: —, Step 3: 40.3 (#157)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| LMArena Longer Query | — | 1326 |
Writing & Preference Step 3 leads
Nova Premier 1.0: 51.0 (#178), Step 3: 54.3 (#151)
| Benchmark | Nova Premier 1.0 | Step 3 |
|---|---|---|
| LMArena Text | — | 1350 |
| LMArena Creative Writing | — | 1321 |
| WildBench | 78.8% | — |
| LMArena Multi-Turn | — | 1341 |
Frequently asked questions
Is Nova Premier 1.0 better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 38.3 on the Noometry Index.
How many benchmarks do Nova Premier 1.0 and Step 3 share?
1 benchmark has published results for both models. Nova Premier 1.0 has 6 scored results on Noometry and Step 3 has 17.