# GPT-4o mini vs Phi-4

> Phi-4 is the stronger model overall, scoring 31.2 to 25.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4o-mini-vs-phi-4
- Last updated: 2026-10-10
- Shared benchmarks: 35

## Summary

- They share 35 benchmarks with published results for both. GPT-4o mini scores higher in 4 categories and Phi-4 in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Phi-4 leads 32.6 to 17.7.
- The biggest single-benchmark swing is GPQA Diamond: 37.7% for GPT-4o mini and 56.1% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.15 / $0.60 for GPT-4o mini.
- Phi-4 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4o mini | Phi-4 |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 25.5 | 31.2 |
| Rank | 343 | 279 |
| Context | 128K | 128K |
| Input $/M | $0.15 | $0.07 |
| Output $/M | $0.60 | $0.14 |
| Weights | Proprietary | Open |

## Coding

- GPT-4o mini: 22.0 (#335)
- Phi-4: 34.4 (#239)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| BigCodeBench Instruct | 46.1% | 45.5% |
| LiveBench Coding | 43.1% | 30.7% |
| LMArena Coding | 1290 | 1231 |
| BigCodeBench Complete | 57.4% | 55.4% |
| Aider Polyglot | 3.6% | — |
| WeirdML | 11.8% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 72.2% | — |

## Agentic & Tool Use

- GPT-4o mini: 27.5 (#101)
- Phi-4: 22.8 (#128)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| BALROG | 17.4% | 11.6% |
| Berkeley Function Calling Leaderboard | — | 28.8% |

## Reasoning

- GPT-4o mini: 8.7 (#347)
- Phi-4: 17.7 (#291)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| Chess Puzzles | 0% | 1% |
| LiveBench Reasoning | 32.8% | 47.8% |
| LMArena Hard Prompts | 1267 | 1220 |
| LiveBench Data Analysis | 50% | 45.2% |
| Epoch Capabilities Index | 126.56 | 130.42 |
| LiveBench | 41.3% | 41.6% |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 10.7% | — |
| Kagi LLM Benchmark | 28.8% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 54.4% | — |
| LMCA | 10.4% | — |
| PIQA | 88.7% | — |

## Math

- GPT-4o mini: 10.4 (#314)
- Phi-4: 20.8 (#285)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.9% | 13.8% |
| LiveBench Math | 36.3% | 42% |
| LMArena Math | 1267 | 1246 |
| MATH Level 5 | 52.6% | 64.9% |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| Omni-MATH | 28% | — |
| GSM8K | 91.3% | — |

## Knowledge

- GPT-4o mini: 17.7 (#284)
- Phi-4: 32.6 (#209)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| GPQA Diamond | 37.7% | 56.1% |
| Confabulations | 37.2% | 29.4% |
| LMArena Expert | 1235 | 1203 |
| MMLU | 81.8% | 84.8% |
| SimpleQA Verified | 8.3% | — |
| MMLU-Pro | 60.3% | — |
| Vectara Hallucination Rate | — | 3.7% |
| GPQA (HELM) | 36.8% | — |
| BoolQ | 88.7% | — |

## Multimodal

- GPT-4o mini: 25.9 (#122)
- Phi-4: —

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |

## Multilingual

- GPT-4o mini: 42.0 (#199)
- Phi-4: 37.2 (#237)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| LMArena Non-English | 1266 | 1197 |
| LMArena Chinese | 1265 | 1212 |
| LMArena French | 1297 | 1224 |
| LMArena German | 1272 | 1222 |
| LMArena Japanese | 1216 | 1158 |
| LMArena Korean | 1195 | 1151 |
| LMArena Russian | 1275 | 1209 |
| LMArena Spanish | 1276 | 1234 |

## Instruction Following

- GPT-4o mini: 61.9 (#239)
- Phi-4: 60.4 (#251)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| LiveBench Instruction Following | 56.8% | 58.4% |
| LMArena Instruction Following | 1258 | 1201 |
| IFEval | 78.2% | — |

## Long Context

- GPT-4o mini: 39.1 (#186)
- Phi-4: 36.9 (#226)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1289 | 1217 |

## Writing & Preference

- GPT-4o mini: 39.5 (#248)
- Phi-4: 40.5 (#244)

| Benchmark | GPT-4o mini | Phi-4 |
|---|---|---|
| LMArena Text | 1286 | 1217 |
| LMArena Creative Writing | 1268 | 1182 |
| Short-Story Creative Writing | 67.2% | 62.6% |
| LMArena Multi-Turn | 1285 | 1206 |
| LiveBench Language | 28.6% | 25.6% |
| EQ-Bench Creative Writing | 873 | — |
| WildBench | 79.1% | — |

## FAQ

### Is GPT-4o mini better than Phi-4?

Phi-4 is the stronger model overall, scoring 31.2 to 25.5 on the Noometry Index.

### Which is cheaper, GPT-4o mini or Phi-4?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; GPT-4o mini lists at $0.15 and $0.60.

### Is GPT-4o mini or Phi-4 better for coding?

Phi-4 scores higher on coding benchmarks: 34.4 versus 22.0 in the Noometry coding category.

### Which has the bigger context window?

Both accept 128K tokens.

### How many benchmarks do GPT-4o mini and Phi-4 share?

35 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Phi-4 has 37.