Model comparison
Gemini 2.5 Flash-Lite vs Phi-4
Gemini 2.5 Flash-Lite is the stronger model overall, scoring 37.0 to 31.2 on the Noometry Index. Phi-4 costs 2.0× less per token, which makes it the better buy when Gemini 2.5 Flash-Lite's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
Google
MicrosoftSummary
- They share 20 benchmarks with published results for both. Gemini 2.5 Flash-Lite scores higher in 7 categories and Phi-4 in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemini 2.5 Flash-Lite leads 38.0 to 20.8.
- The biggest single-benchmark swing is Berkeley Function Calling Leaderboard: 36.9% for Gemini 2.5 Flash-Lite and 28.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.10 / $0.40 for Gemini 2.5 Flash-Lite.
- Gemini 2.5 Flash-Lite accepts more context: 1.05M tokens versus 128K.
- Phi-4 has downloadable open weights; the other is API-only.
Side by side
| Gemini 2.5 Flash-Lite | Phi-4 | |
|---|---|---|
| Provider | Microsoft | |
| Noometry Index | 37.0 | 31.2 |
| Released | 2025-06-17 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 66K | 4K |
| Input $ / M tokens | $0.10 | $0.07 |
| Output $ / M tokens | $0.40 | $0.14 |
| Results tracked | 33 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 38.5 (#173), Phi-4: 34.4 (#239)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Coding | 1373 | 1231 |
| WeirdML | 35.2% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 325.9 | — |
Agentic & Tool Use Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 28.0 (#96), Phi-4: 22.8 (#128)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 36.9% | 28.8% |
| BALROG | — | 11.6% |
Reasoning Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 22.2 (#205), Phi-4: 17.7 (#291)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1377 | 1220 |
| Epoch Capabilities Index | 133.94 | 130.42 |
| Kagi LLM Benchmark | 40.5% | — |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 62.8% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 18.1% | — |
| LiveBench | — | 41.6% |
Math Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 38.0 (#144), Phi-4: 20.8 (#285)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Math | 1373 | 1246 |
| OTIS Mock AIME 2024-2025 | — | 13.8% |
| Omni-MATH | 48% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge Too close to call
Gemini 2.5 Flash-Lite: 32.5 (#210), Phi-4: 32.6 (#209)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| Vectara Hallucination Rate | 3.3% | 3.7% |
| LMArena Expert | 1373 | 1203 |
| GPQA Diamond | — | 56.1% |
| MMLU-Pro | 53.7% | — |
| Confabulations | — | 29.4% |
| GPQA (HELM) | 30.9% | — |
| MMLU | — | 84.8% |
Multimodal Not comparable
Gemini 2.5 Flash-Lite: 29.1 (#114), Phi-4: —
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Vision | 1198 | — |
| VPCT | 30% | — |
Multilingual Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 49.3 (#134), Phi-4: 37.2 (#237)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Non-English | 1369 | 1197 |
| LMArena Chinese | 1404 | 1212 |
| LMArena French | 1388 | 1224 |
| LMArena German | 1389 | 1222 |
| LMArena Japanese | 1359 | 1158 |
| LMArena Korean | 1360 | 1151 |
| LMArena Russian | 1373 | 1209 |
| LMArena Spanish | 1396 | 1234 |
Instruction Following Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 70.0 (#168), Phi-4: 60.4 (#251)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1367 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
| IFEval | 81% | — |
Long Context Phi-4 leads
Gemini 2.5 Flash-Lite: 33.3 (#262), Phi-4: 36.9 (#226)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1373 | 1217 |
| Fiction.LiveBench | 47.2% | — |
Writing & Preference Gemini 2.5 Flash-Lite leads
Gemini 2.5 Flash-Lite: 56.8 (#135), Phi-4: 40.5 (#244)
| Benchmark | Gemini 2.5 Flash-Lite | Phi-4 |
|---|---|---|
| LMArena Text | 1379 | 1217 |
| LMArena Creative Writing | 1367 | 1182 |
| LMArena Multi-Turn | 1366 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| WildBench | 81.8% | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Gemini 2.5 Flash-Lite better than Phi-4?
Gemini 2.5 Flash-Lite is the stronger model overall, scoring 37.0 to 31.2 on the Noometry Index. Phi-4 costs 2.0× less per token, which makes it the better buy when Gemini 2.5 Flash-Lite's lead doesn't matter for your workload.
Which is cheaper, Gemini 2.5 Flash-Lite or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Gemini 2.5 Flash-Lite lists at $0.10 and $0.40.
Is Gemini 2.5 Flash-Lite or Phi-4 better for coding?
Gemini 2.5 Flash-Lite scores higher on coding benchmarks: 38.5 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Gemini 2.5 Flash-Lite does, with 1.05M tokens against 128K.
How many benchmarks do Gemini 2.5 Flash-Lite and Phi-4 share?
20 benchmarks have published results for both models. Gemini 2.5 Flash-Lite has 33 scored results on Noometry and Phi-4 has 37.