Model comparison
Gemini 2.5 Pro vs Phi-4 Mini
Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 30.9 on the Noometry Index. Phi-4 Mini costs 26× less per token, which makes it the better buy when Gemini 2.5 Pro's lead doesn't matter for your workload.
Last verified . 3 shared benchmarks.
Google
MicrosoftSummary
- They share 3 benchmarks with published results for both. Gemini 2.5 Pro scores higher in 3 categories and Phi-4 Mini in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemini 2.5 Pro leads 56.0 to 25.3.
- The biggest single-benchmark swing is SciCode: 42.8% for Gemini 2.5 Pro and 10.8% for Phi-4 Mini.
- Phi-4 Mini is cheaper at $0.075 / $0.30 per million input/output tokens, against $1.25 / $10 for Gemini 2.5 Pro.
- Gemini 2.5 Pro accepts more context: 1.05M tokens versus 128K.
- Phi-4 Mini has downloadable open weights; the other is API-only.
Side by side
| Gemini 2.5 Pro | Phi-4 Mini | |
|---|---|---|
| Provider | Microsoft | |
| Noometry Index | 45.0 | 30.9 |
| Released | 2025-03-25 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 66K | 4K |
| Input $ / M tokens | $1.25 | $0.075 |
| Output $ / M tokens | $10 | $0.30 |
| Results tracked | 78 | 3 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 2.5 Pro leads
Gemini 2.5 Pro: 42.4 (#101), Phi-4 Mini: 28.1 (#317)
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| SciCode | 42.8% | 10.8% |
| SWE-bench Verified | 57.6% | — |
| SWE-bench Verified (bash only) | 53.6% | — |
| Aider Polyglot | 83.1% | — |
| LMArena WebDev | 1227 | — |
| GSO | 3.9% | — |
| WeirdML | 54% | — |
| LiveBench Coding | 85.9% | — |
| LMArena Coding | 1452 | — |
| CadEval | 64% | — |
| ALE-Bench | 785.52 | — |
| AlgoTune | 1.51 | — |
Agentic & Tool Use Not comparable
Gemini 2.5 Pro: 29.2 (#88), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| Terminal-Bench | 32.6% | — |
| GDPval | 23.3% | — |
| Remote Labor Index | 0.8% | — |
| TheAgentCompany | 30.3% | — |
| τ²-bench Banking | 13.7% | — |
| DeepResearch Bench | 42.8% | — |
| BALROG | 43.3% | — |
| LMArena Search | 1142 | — |
| METR Time Horizons | 55.4% | — |
| Vending-Bench 2 | 573.64 | — |
Reasoning Gemini 2.5 Pro leads
Gemini 2.5 Pro: 28.8 (#99), Phi-4 Mini: 22.4 (#195)
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| CritPt | 2% | 0% |
| ARC-AGI-2 | 4.9% | — |
| SimpleBench | 62.4% | — |
| Kagi LLM Benchmark | 70.3% | — |
| ARC-AGI-1 | 41% | — |
| Chess Puzzles | 20% | — |
| EnigmaEval | 5.6% | — |
| LiveBench Reasoning | 89.8% | — |
| LMArena Hard Prompts | 1455 | — |
| DTBench | 82.4% | — |
| LiveBench Data Analysis | 79.9% | — |
| LMCA | 34.8% | — |
| Epoch Capabilities Index | 145.32 | — |
| ForecastBench | 61.3 | — |
| LiveBench | 82.3% | — |
Math Not comparable
Gemini 2.5 Pro: 32.5 (#213), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 24.6% | — |
| FrontierMath Tier 4 | 0% | — |
| OTIS Mock AIME 2024-2025 | 84.7% | — |
| Omni-MATH | 41.6% | — |
| LiveBench Math | 90.2% | — |
| LMArena Math | 1450 | — |
| MATH Level 5 | 95.9% | — |
| FrontierMath (Feb 2025 set) | 14.1% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge Gemini 2.5 Pro leads
Gemini 2.5 Pro: 56.0 (#46), Phi-4 Mini: 25.3 (#262)
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| Vectara Hallucination Rate | 7% | 23.5% |
| GPQA Diamond | 85.3% | — |
| Humanity's Last Exam | 21.6% | — |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.6% | — |
| GPQA (HELM) | 74.9% | — |
| LMArena Expert | 1452 | — |
Multimodal Not comparable
Gemini 2.5 Pro: 45.2 (#18), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| LMArena Vision | 1263 | — |
| GeoBench | 86% | — |
| VPCT | 48% | — |
| LMArena Document | 1421 | — |
| SpatialViz-Bench | 44.7% | — |
Multilingual Not comparable
Gemini 2.5 Pro: 55.3 (#31), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| LMArena Non-English | 1451 | — |
| LMArena Chinese | 1507 | — |
| LMArena French | 1472 | — |
| LMArena German | 1487 | — |
| LMArena Japanese | 1461 | — |
| LMArena Korean | 1434 | — |
| LMArena Russian | 1461 | — |
| LMArena Spanish | 1473 | — |
Instruction Following Not comparable
Gemini 2.5 Pro: 75.0 (#75), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| LiveBench Instruction Following | 80.6% | — |
| IFEval | 84% | — |
| LMArena Instruction Following | 1437 | — |
Long Context Not comparable
Gemini 2.5 Pro: 59.8 (#5), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| Fiction.LiveBench | 91.7% | — |
| LMArena Longer Query | 1449 | — |
Writing & Preference Not comparable
Gemini 2.5 Pro: 63.7 (#62), Phi-4 Mini: —
| Benchmark | Gemini 2.5 Pro | Phi-4 Mini |
|---|---|---|
| LMArena Text | 1458 | — |
| LMArena Creative Writing | 1454 | — |
| Short-Story Creative Writing | 83.8% | — |
| EQ-Bench Creative Writing | 1421 | — |
| WildBench | 85.7% | — |
| LMArena Multi-Turn | 1453 | — |
| LiveBench Language | 67.8% | — |
Frequently asked questions
Is Gemini 2.5 Pro better than Phi-4 Mini?
Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 30.9 on the Noometry Index. Phi-4 Mini costs 26× less per token, which makes it the better buy when Gemini 2.5 Pro's lead doesn't matter for your workload.
Which is cheaper, Gemini 2.5 Pro or Phi-4 Mini?
Phi-4 Mini is cheaper. It lists at $0.075 per million input tokens and $0.30 per million output tokens; Gemini 2.5 Pro lists at $1.25 and $10.
Is Gemini 2.5 Pro or Phi-4 Mini better for coding?
Gemini 2.5 Pro scores higher on coding benchmarks: 42.4 versus 28.1 in the Noometry coding category.
Which has the bigger context window?
Gemini 2.5 Pro does, with 1.05M tokens against 128K.
How many benchmarks do Gemini 2.5 Pro and Phi-4 Mini share?
3 benchmarks have published results for both models. Gemini 2.5 Pro has 78 scored results on Noometry and Phi-4 Mini has 3.