Model comparison
Gemini 3.1 Flash Lite vs Phi-4
Gemini 3.1 Flash Lite is the stronger model overall, scoring 40.8 to 31.2 on the Noometry Index. Phi-4 costs 6.4× less per token, which makes it the better buy when Gemini 3.1 Flash Lite's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
Google
MicrosoftSummary
- They share 22 benchmarks with published results for both. Gemini 3.1 Flash Lite scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemini 3.1 Flash Lite leads 60.9 to 40.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 80% for Gemini 3.1 Flash Lite and 13.8% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.25 / $1.50 for Gemini 3.1 Flash Lite.
- Gemini 3.1 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 3.1 Flash Lite | Phi-4 | |
|---|---|---|
| Provider | Microsoft | |
| Noometry Index | 40.8 | 31.2 |
| Released | 2026-03-03 | 2024-12-11 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 66K | 4K |
| Input $ / M tokens | $0.25 | $0.07 |
| Output $ / M tokens | $1.50 | $0.14 |
| Results tracked | 38 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 37.8 (#188), Phi-4: 34.4 (#239)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Coding | 1400 | 1231 |
| LMArena WebDev | 1256 | — |
| SciCode | 41.9% | — |
| WeirdML | 52.2% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 797.73 | — |
Agentic & Tool Use Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 30.2 (#79), Phi-4: 22.8 (#128)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| DeepResearch Bench | 37.3% | — |
| BALROG | — | 11.6% |
Reasoning Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 22.9 (#186), Phi-4: 17.7 (#291)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| Chess Puzzles | 25% | 1% |
| LMArena Hard Prompts | 1407 | 1220 |
| Epoch Capabilities Index | 144.47 | 130.42 |
| Kagi LLM Benchmark | 67.2% | — |
| NYT Connections (extended) | 8.2% | — |
| CritPt | 1.1% | — |
| EnigmaEval | 3% | — |
| Thematic Generalization | 63.3% | — |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 76.8% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 35% | — |
| ForecastBench | 54.4 | — |
| LiveBench | — | 41.6% |
Math Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 40.7 (#90), Phi-4: 20.8 (#285)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 80% | 13.8% |
| LMArena Math | 1428 | 1246 |
| FrontierMath (Tiers 1-3) | 27.7% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 41.9 (#104), Phi-4: 32.6 (#209)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| GPQA Diamond | 81.8% | 56.1% |
| Vectara Hallucination Rate | 8.2% | 3.7% |
| LMArena Expert | 1398 | 1203 |
| Humanity's Last Exam | 8.6% | — |
| Confabulations | — | 29.4% |
| MMLU | — | 84.8% |
Multimodal Not comparable
Gemini 3.1 Flash Lite: 39.4 (#60), Phi-4: —
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Vision | 1240 | — |
Multilingual Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 52.3 (#86), Phi-4: 37.2 (#237)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Non-English | 1411 | 1197 |
| LMArena Chinese | 1461 | 1212 |
| LMArena French | 1424 | 1224 |
| LMArena German | 1429 | 1222 |
| LMArena Japanese | 1413 | 1158 |
| LMArena Korean | 1392 | 1151 |
| LMArena Russian | 1420 | 1209 |
| LMArena Spanish | 1421 | 1234 |
Instruction Following Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 72.7 (#131), Phi-4: 60.4 (#251)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1377 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 42.5 (#122), Phi-4: 36.9 (#226)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1394 | 1217 |
Writing & Preference Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 60.9 (#94), Phi-4: 40.5 (#244)
| Benchmark | Gemini 3.1 Flash Lite | Phi-4 |
|---|---|---|
| LMArena Text | 1416 | 1217 |
| LMArena Creative Writing | 1401 | 1182 |
| LMArena Multi-Turn | 1417 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is Gemini 3.1 Flash Lite better than Phi-4?
Gemini 3.1 Flash Lite is the stronger model overall, scoring 40.8 to 31.2 on the Noometry Index. Phi-4 costs 6.4× less per token, which makes it the better buy when Gemini 3.1 Flash Lite's lead doesn't matter for your workload.
Which is cheaper, Gemini 3.1 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 3.1 Flash Lite lists at $0.25 and $1.50.
Is Gemini 3.1 Flash Lite or Phi-4 better for coding?
Gemini 3.1 Flash Lite scores higher on coding benchmarks: 37.8 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.1 Flash Lite does, with 1.05M tokens against 128K.
How many benchmarks do Gemini 3.1 Flash Lite and Phi-4 share?
22 benchmarks have published results for both models. Gemini 3.1 Flash Lite has 38 scored results on Noometry and Phi-4 has 37.