Model comparison
Deepseek Coder v2 vs Gemma 2 27B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index.
Last verified . 19 shared benchmarks.
DeepSeek
GoogleSummary
- They share 19 benchmarks with published results for both. Deepseek Coder v2 scores higher in 5 categories and Gemma 2 27B in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Deepseek Coder v2 leads 34.9 to 10.7.
- The biggest single-benchmark swing is BigCodeBench Complete: 59.7% for Deepseek Coder v2 and 52.5% for Gemma 2 27B.
Side by side
| Deepseek Coder v2 | Gemma 2 27B | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 35.9 | 29.4 |
| Released | 2024-06-17 | 2024-06-24 |
| Weights | Open | Open |
| Context window | — | 8K |
| Max output | — | 2K |
| Input $ / M tokens | — | $0.65 |
| Output $ / M tokens | — | $0.65 |
| Results tracked | 24 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 38.1 (#183), Gemma 2 27B: 34.1 (#246)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| BigCodeBench Instruct | 48.2% | 42.8% |
| LMArena Coding | 1251 | 1211 |
| BigCodeBench Complete | 59.7% | 52.5% |
| LiveBench Coding | — | 36% |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Gemma 2 27B: 15.3 (#315)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1198 |
| LiveBench Reasoning | — | 28.1% |
| DTBench | — | 48% |
| LiveBench Data Analysis | — | 47.9% |
| LMCA | — | 7.1% |
| Epoch Capabilities Index | — | 122.08 |
| LiveBench | — | 38.2% |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Gemma 2 27B: 10.7 (#311)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Math | 1241 | 1212 |
| OTIS Mock AIME 2024-2025 | — | 1.4% |
| LiveBench Math | — | 26.5% |
| MATH Level 5 | — | 27.9% |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Gemma 2 27B: 19.0 (#280)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Expert | 1181 | 1172 |
| GPQA Diamond | — | 36.5% |
| Confabulations | — | 27.1% |
| ARC (AI2) Challenge | 64.3% | — |
| MMLU | — | 75.7% |
Multilingual Gemma 2 27B leads
Deepseek Coder v2: 36.3 (#240), Gemma 2 27B: 38.6 (#226)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Non-English | 1182 | 1217 |
| LMArena Chinese | 1201 | 1221 |
| LMArena French | 1185 | 1247 |
| LMArena German | 1164 | 1209 |
| LMArena Japanese | 1126 | 1175 |
| LMArena Korean | 1104 | 1174 |
| LMArena Russian | 1188 | 1234 |
| LMArena Spanish | 1153 | 1228 |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.7 (#242), Gemma 2 27B: 60.5 (#249)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1206 |
| LiveBench Instruction Following | — | 58.1% |
Long Context Too close to call
Deepseek Coder v2: 37.0 (#224), Gemma 2 27B: 37.3 (#218)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Longer Query | 1219 | 1231 |
Writing & Preference Gemma 2 27B leads
Deepseek Coder v2: 38.2 (#253), Gemma 2 27B: 44.2 (#225)
| Benchmark | Deepseek Coder v2 | Gemma 2 27B |
|---|---|---|
| LMArena Text | 1191 | 1231 |
| LMArena Creative Writing | 1120 | 1241 |
| LMArena Multi-Turn | 1177 | 1224 |
| LiveBench Language | — | 32.6% |
Frequently asked questions
Is Deepseek Coder v2 better than Gemma 2 27B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index.
Is Deepseek Coder v2 or Gemma 2 27B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 34.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Gemma 2 27B share?
19 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Gemma 2 27B has 34.