Model comparison
Gemma 2 9B vs Phi 3 Mini 128k Instruct
Phi 3 Mini 128k Instruct is the stronger model overall, scoring 29.7 to 25.9 on the Noometry Index.
Last verified . 19 shared benchmarks.
Google
MicrosoftSummary
- They share 19 benchmarks with published results for both. Gemma 2 9B scores higher in 5 categories and Phi 3 Mini 128k Instruct in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Phi 3 Mini 128k Instruct leads 31.6 to 9.9.
- The biggest single-benchmark swing is BigCodeBench Instruct: 34.7% for Gemma 2 9B and 29.6% for Phi 3 Mini 128k Instruct.
Side by side
| Gemma 2 9B | Phi 3 Mini 128k Instruct | |
|---|---|---|
| Provider | Microsoft | |
| Noometry Index | 25.9 | 29.7 |
| Released | 2024-06-24 | 2024-04-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 2 9B: 29.4 (#304), Phi 3 Mini 128k Instruct: 28.8 (#312)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| BigCodeBench Instruct | 34.7% | 29.6% |
| LMArena Coding | 1173 | 1039 |
| BigCodeBench Complete | 40.6% | 40.6% |
| LiveBench Coding | 22.5% | — |
Reasoning Phi 3 Mini 128k Instruct leads
Gemma 2 9B: 15.9 (#309), Phi 3 Mini 128k Instruct: 19.6 (#256)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Hard Prompts | 1171 | 1028 |
| LiveBench Reasoning | 15.2% | — |
| LiveBench Data Analysis | 36.4% | — |
| Epoch Capabilities Index | 119.83 | — |
| LiveBench | 28.7% | — |
| PIQA | 83.7% | — |
Math Phi 3 Mini 128k Instruct leads
Gemma 2 9B: 9.9 (#318), Phi 3 Mini 128k Instruct: 31.6 (#222)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Math | 1183 | 1089 |
| OTIS Mock AIME 2024-2025 | 0.6% | — |
| LiveBench Math | 19.8% | — |
| MATH Level 5 | 21% | — |
| GSM8K | 84.9% | — |
Knowledge Phi 3 Mini 128k Instruct leads
Gemma 2 9B: 9.7 (#305), Phi 3 Mini 128k Instruct: 26.8 (#254)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Expert | 1147 | 984 |
| GPQA Diamond | 27.5% | — |
| BoolQ | 85.7% | — |
| MMLU | 72.1% | — |
Multilingual Gemma 2 9B leads
Gemma 2 9B: 36.6 (#238), Phi 3 Mini 128k Instruct: 25.2 (#285)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Non-English | 1188 | 1000 |
| LMArena Chinese | 1185 | 1016 |
| LMArena French | 1190 | 1039 |
| LMArena German | 1186 | 1006 |
| LMArena Japanese | 1144 | 899 |
| LMArena Korean | 1137 | 856 |
| LMArena Russian | 1200 | 1004 |
| LMArena Spanish | 1200 | 1059 |
Instruction Following Gemma 2 9B leads
Gemma 2 9B: 57.6 (#269), Phi 3 Mini 128k Instruct: 51.8 (#294)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Instruction Following | 1178 | 1023 |
| LiveBench Instruction Following | 52.6% | — |
Long Context Gemma 2 9B leads
Gemma 2 9B: 36.3 (#233), Phi 3 Mini 128k Instruct: 30.4 (#289)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Longer Query | 1197 | 996 |
Writing & Preference Gemma 2 9B leads
Gemma 2 9B: 32.1 (#281), Phi 3 Mini 128k Instruct: 27.1 (#301)
| Benchmark | Gemma 2 9B | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Text | 1207 | 1050 |
| LMArena Creative Writing | 1206 | 1024 |
| LMArena Multi-Turn | 1193 | 989 |
| EQ-Bench Creative Writing | 841 | — |
| LiveBench Language | 25.5% | — |
Frequently asked questions
Is Gemma 2 9B better than Phi 3 Mini 128k Instruct?
Phi 3 Mini 128k Instruct is the stronger model overall, scoring 29.7 to 25.9 on the Noometry Index.
Is Gemma 2 9B or Phi 3 Mini 128k Instruct better for coding?
They score almost the same on coding (29.4 vs 28.8); test both on your own repository before choosing.
How many benchmarks do Gemma 2 9B and Phi 3 Mini 128k Instruct share?
19 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and Phi 3 Mini 128k Instruct has 19.