Model comparison
Phi-4 vs Step 3
Step 3 is the stronger model overall, scoring 40.5 to 31.2 on the Noometry Index.
Last verified . 15 shared benchmarks.
Microsoft
StepFunSummary
- They share 15 benchmarks with published results for both. Phi-4 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 20.8.
Side by side
| Phi-4 | Step 3 | |
|---|---|---|
| Provider | Microsoft | StepFun |
| Noometry Index | 31.2 | 40.5 |
| Released | 2024-12-11 | — |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.07 | — |
| Output $ / M tokens | $0.14 | — |
| Results tracked | 37 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 3 leads
Phi-4: 34.4 (#239), Step 3: 40.1 (#147)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Coding | 1231 | 1367 |
| BigCodeBench Instruct | 45.5% | — |
| LiveBench Coding | 30.7% | — |
| BigCodeBench Complete | 55.4% | — |
Agentic & Tool Use Not comparable
Phi-4: 22.8 (#128), Step 3: —
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |
Reasoning Step 3 leads
Phi-4: 17.7 (#291), Step 3: 28.4 (#105)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1355 |
| Kagi LLM Benchmark | — | 62.3% |
| Chess Puzzles | 1% | — |
| LiveBench Reasoning | 47.8% | — |
| LiveBench Data Analysis | 45.2% | — |
| Epoch Capabilities Index | 130.42 | — |
| LiveBench | 41.6% | — |
Math Step 3 leads
Phi-4: 20.8 (#285), Step 3: 37.6 (#148)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Math | 1246 | 1366 |
| OTIS Mock AIME 2024-2025 | 13.8% | — |
| LiveBench Math | 42% | — |
| MATH Level 5 | 64.9% | — |
Knowledge Step 3 leads
Phi-4: 32.6 (#209), Step 3: 36.8 (#164)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Expert | 1203 | 1333 |
| GPQA Diamond | 56.1% | — |
| Confabulations | 29.4% | — |
| Vectara Hallucination Rate | 3.7% | — |
| MMLU | 84.8% | — |
Multimodal Not comparable
Phi-4: —, Step 3: 35.5 (#86)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Vision | — | 1177 |
Multilingual Step 3 leads
Phi-4: 37.2 (#237), Step 3: 46.3 (#159)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Non-English | 1197 | 1327 |
| LMArena Chinese | 1212 | 1397 |
| LMArena German | 1222 | 1371 |
| LMArena Korean | 1151 | 1269 |
| LMArena Russian | 1209 | 1331 |
| LMArena Spanish | 1234 | 1371 |
| LMArena French | 1224 | — |
| LMArena Japanese | 1158 | — |
Instruction Following Step 3 leads
Phi-4: 60.4 (#251), Step 3: 70.4 (#164)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Instruction Following | 1201 | 1332 |
| LiveBench Instruction Following | 58.4% | — |
Long Context Step 3 leads
Phi-4: 36.9 (#226), Step 3: 40.3 (#157)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Longer Query | 1217 | 1326 |
Writing & Preference Step 3 leads
Phi-4: 40.5 (#244), Step 3: 54.3 (#151)
| Benchmark | Phi-4 | Step 3 |
|---|---|---|
| LMArena Text | 1217 | 1350 |
| LMArena Creative Writing | 1182 | 1321 |
| LMArena Multi-Turn | 1206 | 1341 |
| Short-Story Creative Writing | 62.6% | — |
| LiveBench Language | 25.6% | — |
Frequently asked questions
Is Phi-4 better than Step 3?
Step 3 is the stronger model overall, scoring 40.5 to 31.2 on the Noometry Index.
Is Phi-4 or Step 3 better for coding?
Step 3 scores higher on coding benchmarks: 40.1 versus 34.4 in the Noometry coding category.
How many benchmarks do Phi-4 and Step 3 share?
15 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Step 3 has 17.