Model comparison
GPT-5 Pro vs Phi-4
GPT-5 Pro is the stronger model overall, scoring 46.4 to 31.2 on the Noometry Index. Phi-4 costs 471× less per token, which makes it the better buy when GPT-5 Pro's lead doesn't matter for your workload.
Last verified . 1 shared benchmarks.
MicrosoftSummary
- They share 1 benchmark with published results for both. GPT-5 Pro scores higher in 4 categories and Phi-4 in 0 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 Pro leads 48.5 to 20.8.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $15 / $120 for GPT-5 Pro.
- GPT-5 Pro accepts more context: 400K tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Pro | Phi-4 | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 46.4 | 31.2 |
| Released | 2025-10-06 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 400K | 128K |
| Max output | 272K | 4K |
| Input $ / M tokens | $15 | $0.07 |
| Output $ / M tokens | $120 | $0.14 |
| Results tracked | 12 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Pro leads
GPT-5 Pro: 44.0 (#80), Phi-4: 34.4 (#239)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| WeirdML | 60.4% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| LMArena Coding | — | 1231 |
| BigCodeBench Complete | — | 55.4% |
| AlgoTune | 1.31 | — |
Agentic & Tool Use Not comparable
GPT-5 Pro: —, Phi-4: 22.8 (#128)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |
Reasoning GPT-5 Pro leads
GPT-5 Pro: 38.9 (#62), Phi-4: 17.7 (#291)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| Epoch Capabilities Index | 150.28 | 130.42 |
| ARC-AGI-2 | 18.3% | — |
| SimpleBench | 61.6% | — |
| Kagi LLM Benchmark | 76.8% | — |
| ARC-AGI-1 | 70.2% | — |
| Chess Puzzles | — | 1% |
| EnigmaEval | 18.8% | — |
| LiveBench Reasoning | — | 47.8% |
| LMArena Hard Prompts | — | 1220 |
| LiveBench Data Analysis | — | 45.2% |
| LiveBench | — | 41.6% |
Math GPT-5 Pro leads
GPT-5 Pro: 48.5 (#63), Phi-4: 20.8 (#285)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 55.8% | — |
| FrontierMath Tier 4 | 19.5% | — |
| OTIS Mock AIME 2024-2025 | — | 13.8% |
| LiveBench Math | — | 42% |
| LMArena Math | — | 1246 |
| MATH Level 5 | — | 64.9% |
| FrontierMath Tier 4 (v1) | 14.6% | — |
Knowledge GPT-5 Pro leads
GPT-5 Pro: 56.7 (#42), Phi-4: 32.6 (#209)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| GPQA Diamond | — | 56.1% |
| Humanity's Last Exam | 31.6% | — |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| LMArena Expert | — | 1203 |
| MMLU | — | 84.8% |
Multilingual Not comparable
GPT-5 Pro: —, Phi-4: 37.2 (#237)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| LMArena Non-English | — | 1197 |
| LMArena Chinese | — | 1212 |
| LMArena French | — | 1224 |
| LMArena German | — | 1222 |
| LMArena Japanese | — | 1158 |
| LMArena Korean | — | 1151 |
| LMArena Russian | — | 1209 |
| LMArena Spanish | — | 1234 |
Instruction Following Not comparable
GPT-5 Pro: —, Phi-4: 60.4 (#251)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| LiveBench Instruction Following | — | 58.4% |
| LMArena Instruction Following | — | 1201 |
Long Context Not comparable
GPT-5 Pro: —, Phi-4: 36.9 (#226)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| LMArena Longer Query | — | 1217 |
Writing & Preference Not comparable
GPT-5 Pro: —, Phi-4: 40.5 (#244)
| Benchmark | GPT-5 Pro | Phi-4 |
|---|---|---|
| LMArena Text | — | 1217 |
| LMArena Creative Writing | — | 1182 |
| Short-Story Creative Writing | — | 62.6% |
| LMArena Multi-Turn | — | 1206 |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is GPT-5 Pro better than Phi-4?
GPT-5 Pro is the stronger model overall, scoring 46.4 to 31.2 on the Noometry Index. Phi-4 costs 471× less per token, which makes it the better buy when GPT-5 Pro's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Pro or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; GPT-5 Pro lists at $15 and $120.
Is GPT-5 Pro or Phi-4 better for coding?
GPT-5 Pro scores higher on coding benchmarks: 44.0 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Pro does, with 400K tokens against 128K.
How many benchmarks do GPT-5 Pro and Phi-4 share?
1 benchmark has published results for both models. GPT-5 Pro has 12 scored results on Noometry and Phi-4 has 37.