Model comparison
o1-mini vs Phi-4
o1-mini is the stronger model overall, scoring 33.7 to 30.9 on the Noometry Index.
Last verified . 30 shared benchmarks.
MicrosoftSummary
- They share 30 benchmarks with published results for both. o1-mini scores higher in 8 categories and Phi-4 in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where o1-mini leads 33.7 to 19.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 46.9% for o1-mini and 13.8% for Phi-4.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| o1-mini | Phi-4 | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 33.7 | 30.9 |
| Released | 2024-09-12 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.07 |
| Output $ / M tokens | — | $0.14 |
| Results tracked | 39 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o1-mini leads
o1-mini: 35.1 (#218), Phi-4: 33.8 (#237)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LiveBench Coding | 48% | 30.7% |
| LMArena Coding | 1362 | 1231 |
| Aider Polyglot | 32.9% | — |
| WeirdML | 36.3% | — |
| BigCodeBench Instruct | — | 45.5% |
| BigCodeBench Complete | — | 55.4% |
| HumanEval+ | 89% | — |
| MBPP+ | 78.8% | — |
Agentic & Tool Use o1-mini leads
o1-mini: 24.3 (#112), Phi-4: 22.6 (#120)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| Cybench | 10% | — |
| BALROG | — | 11.6% |
Reasoning Phi-4 leads
o1-mini: 8.9 (#340), Phi-4: 17.0 (#287)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LiveBench Reasoning | 72.3% | 47.8% |
| LMArena Hard Prompts | 1333 | 1220 |
| LiveBench Data Analysis | 57.9% | 45.2% |
| Epoch Capabilities Index | 135.82 | 130.43 |
| LiveBench | 57.8% | 41.6% |
| ARC-AGI-2 | 0.8% | — |
| SimpleBench | 18.1% | — |
| ARC-AGI-1 | 14% | — |
| Chess Puzzles | — | 1% |
Math o1-mini leads
o1-mini: 33.7 (#168), Phi-4: 19.6 (#279)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 46.9% | 13.8% |
| LiveBench Math | 62% | 42% |
| LMArena Math | 1358 | 1246 |
| MATH Level 5 | 89.2% | 64.9% |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge o1-mini leads
o1-mini: 34.8 (#180), Phi-4: 32.5 (#204)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| GPQA Diamond | 62.4% | 56.1% |
| Confabulations | 18.6% | 29.4% |
| LMArena Expert | 1316 | 1203 |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |
Multilingual o1-mini leads
o1-mini: 43.8 (#177), Phi-4: 37.4 (#232)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LMArena Non-English | 1289 | 1197 |
| LMArena Chinese | 1314 | 1211 |
| LMArena French | 1292 | 1223 |
| LMArena German | 1278 | 1222 |
| LMArena Japanese | 1245 | 1158 |
| LMArena Korean | 1223 | 1151 |
| LMArena Russian | 1283 | 1209 |
| LMArena Spanish | 1302 | 1234 |
Instruction Following o1-mini leads
o1-mini: 67.0 (#200), Phi-4: 60.6 (#246)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LiveBench Instruction Following | 65.4% | 58.4% |
| LMArena Instruction Following | 1304 | 1201 |
Long Context o1-mini leads
o1-mini: 40.1 (#156), Phi-4: 36.9 (#221)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1321 | 1217 |
Writing & Preference o1-mini leads
o1-mini: 48.6 (#196), Phi-4: 40.7 (#239)
| Benchmark | o1-mini | Phi-4 |
|---|---|---|
| LMArena Text | 1317 | 1217 |
| LMArena Creative Writing | 1244 | 1182 |
| Short-Story Creative Writing | 64.9% | 62.6% |
| LMArena Multi-Turn | 1314 | 1206 |
| LiveBench Language | 40.9% | 25.6% |
Frequently asked questions
Is o1-mini better than Phi-4?
o1-mini is the stronger model overall, scoring 33.7 to 30.9 on the Noometry Index.
Is o1-mini or Phi-4 better for coding?
o1-mini scores higher on coding benchmarks: 35.1 versus 33.8 in the Noometry coding category.
How many benchmarks do o1-mini and Phi-4 share?
30 benchmarks have published results for both models. o1-mini has 39 scored results on Noometry and Phi-4 has 37.