Model comparison
Phi-4 vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 31.2 on the Noometry Index. Phi-4 costs 16× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Last verified . 13 shared benchmarks.
Microsoft
StepFunSummary
- They share 13 benchmarks with published results for both. Phi-4 scores higher in 0 categories and Step 5 Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Step 5 Preview leads 46.1 to 20.8.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| Phi-4 | Step 5 Preview | |
|---|---|---|
| Provider | Microsoft | StepFun |
| Noometry Index | 31.2 | 47.9 |
| Released | 2024-12-11 | 2026-09-16 |
| Weights | Open | Proprietary |
| Context window | 128K | 1.02M |
| Max output | 4K | 66K |
| Input $ / M tokens | $0.07 | $1 |
| Output $ / M tokens | $0.14 | $2.70 |
| Results tracked | 37 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
Phi-4: 34.4 (#239), Step 5 Preview: 51.8 (#36)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Coding | 1231 | 1480 |
| LMArena WebDev | — | 1564 |
| SciCode | — | 58.9% |
| BigCodeBench Instruct | 45.5% | — |
| LiveBench Coding | 30.7% | — |
| BigCodeBench Complete | 55.4% | — |
Agentic & Tool Use Not comparable
Phi-4: 22.8 (#128), Step 5 Preview: —
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |
Reasoning Step 5 Preview leads
Phi-4: 17.7 (#291), Step 5 Preview: 40.0 (#57)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Hard Prompts | 1220 | 1465 |
| CritPt | — | 20.9% |
| Chess Puzzles | 1% | — |
| LiveBench Reasoning | 47.8% | — |
| LiveBench Data Analysis | 45.2% | — |
| Epoch Capabilities Index | 130.42 | — |
| LiveBench | 41.6% | — |
Math Step 5 Preview leads
Phi-4: 20.8 (#285), Step 5 Preview: 46.1 (#72)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Math | 1246 | 1470 |
| OTIS Mock AIME 2024-2025 | 13.8% | — |
| ProofBench | — | 42% |
| LiveBench Math | 42% | — |
| MATH Level 5 | 64.9% | — |
Knowledge Step 5 Preview leads
Phi-4: 32.6 (#209), Step 5 Preview: 41.2 (#112)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1203 | 1470 |
| GPQA Diamond | 56.1% | — |
| Confabulations | 29.4% | — |
| Vectara Hallucination Rate | 3.7% | — |
| MMLU | 84.8% | — |
Multimodal Not comparable
Phi-4: —, Step 5 Preview: 41.0 (#41)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Step 5 Preview leads
Phi-4: 37.2 (#237), Step 5 Preview: 53.8 (#54)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1197 | 1432 |
| LMArena Chinese | 1212 | 1519 |
| LMArena Russian | 1209 | 1436 |
| LMArena Spanish | 1234 | 1439 |
| LMArena French | 1224 | — |
| LMArena German | 1222 | — |
| LMArena Japanese | 1158 | — |
| LMArena Korean | 1151 | — |
Instruction Following Step 5 Preview leads
Phi-4: 60.4 (#251), Step 5 Preview: 76.0 (#50)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1201 | 1444 |
| LiveBench Instruction Following | 58.4% | — |
Long Context Step 5 Preview leads
Phi-4: 36.9 (#226), Step 5 Preview: 44.5 (#64)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1217 | 1455 |
Writing & Preference Step 5 Preview leads
Phi-4: 40.5 (#244), Step 5 Preview: 62.8 (#71)
| Benchmark | Phi-4 | Step 5 Preview |
|---|---|---|
| LMArena Text | 1217 | 1442 |
| LMArena Creative Writing | 1182 | 1410 |
| LMArena Multi-Turn | 1206 | 1447 |
| Short-Story Creative Writing | 62.6% | — |
| LiveBench Language | 25.6% | — |
Frequently asked questions
Is Phi-4 better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 31.2 on the Noometry Index. Phi-4 costs 16× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, Phi-4 or Step 5 Preview?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is Phi-4 or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 128K.
How many benchmarks do Phi-4 and Step 5 Preview share?
13 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and Step 5 Preview has 18.