Model comparison
Gemma 2 9B vs Phi-4
Phi-4 is the stronger model overall, scoring 31.2 to 25.9 on the Noometry Index.
Last verified . 31 shared benchmarks.
Google
MicrosoftSummary
- They share 31 benchmarks with published results for both. Gemma 2 9B scores higher in 0 categories and Phi-4 in 8 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Phi-4 leads 32.6 to 9.7.
- The biggest single-benchmark swing is MATH Level 5: 21% for Gemma 2 9B and 64.9% for Phi-4.
Side by side
| Gemma 2 9B | Phi-4 | |
|---|---|---|
| Provider | Microsoft | |
| Noometry Index | 25.9 | 31.2 |
| Released | 2024-06-24 | 2024-12-11 |
| Weights | Open | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.07 |
| Output $ / M tokens | — | $0.14 |
| Results tracked | 35 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Phi-4 leads
Gemma 2 9B: 29.4 (#304), Phi-4: 34.4 (#239)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| BigCodeBench Instruct | 34.7% | 45.5% |
| LiveBench Coding | 22.5% | 30.7% |
| LMArena Coding | 1173 | 1231 |
| BigCodeBench Complete | 40.6% | 55.4% |
Agentic & Tool Use Not comparable
Gemma 2 9B: —, Phi-4: 22.8 (#128)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |
Reasoning Phi-4 leads
Gemma 2 9B: 15.9 (#309), Phi-4: 17.7 (#291)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| LiveBench Reasoning | 15.2% | 47.8% |
| LMArena Hard Prompts | 1171 | 1220 |
| LiveBench Data Analysis | 36.4% | 45.2% |
| Epoch Capabilities Index | 119.83 | 130.42 |
| LiveBench | 28.7% | 41.6% |
| Chess Puzzles | — | 1% |
| PIQA | 83.7% | — |
Math Phi-4 leads
Gemma 2 9B: 9.9 (#318), Phi-4: 20.8 (#285)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.6% | 13.8% |
| LiveBench Math | 19.8% | 42% |
| LMArena Math | 1183 | 1246 |
| MATH Level 5 | 21% | 64.9% |
| GSM8K | 84.9% | — |
Knowledge Phi-4 leads
Gemma 2 9B: 9.7 (#305), Phi-4: 32.6 (#209)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| GPQA Diamond | 27.5% | 56.1% |
| LMArena Expert | 1147 | 1203 |
| MMLU | 72.1% | 84.8% |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| BoolQ | 85.7% | — |
Multilingual Too close to call
Gemma 2 9B: 36.6 (#238), Phi-4: 37.2 (#237)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| LMArena Non-English | 1188 | 1197 |
| LMArena Chinese | 1185 | 1212 |
| LMArena French | 1190 | 1224 |
| LMArena German | 1186 | 1222 |
| LMArena Japanese | 1144 | 1158 |
| LMArena Korean | 1137 | 1151 |
| LMArena Russian | 1200 | 1209 |
| LMArena Spanish | 1200 | 1234 |
Instruction Following Phi-4 leads
Gemma 2 9B: 57.6 (#269), Phi-4: 60.4 (#251)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| LiveBench Instruction Following | 52.6% | 58.4% |
| LMArena Instruction Following | 1178 | 1201 |
Long Context Too close to call
Gemma 2 9B: 36.3 (#233), Phi-4: 36.9 (#226)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1197 | 1217 |
Writing & Preference Phi-4 leads
Gemma 2 9B: 32.1 (#281), Phi-4: 40.5 (#244)
| Benchmark | Gemma 2 9B | Phi-4 |
|---|---|---|
| LMArena Text | 1207 | 1217 |
| LMArena Creative Writing | 1206 | 1182 |
| LMArena Multi-Turn | 1193 | 1206 |
| LiveBench Language | 25.5% | 25.6% |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 841 | — |
Frequently asked questions
Is Gemma 2 9B better than Phi-4?
Phi-4 is the stronger model overall, scoring 31.2 to 25.9 on the Noometry Index.
Is Gemma 2 9B or Phi-4 better for coding?
Phi-4 scores higher on coding benchmarks: 34.4 versus 29.4 in the Noometry coding category.
How many benchmarks do Gemma 2 9B and Phi-4 share?
31 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and Phi-4 has 37.