Model comparison
DeepSeek-V2.5 (Sep 2024) vs Gemma 2 27B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 29.2 on the Noometry Index.
Last verified . 19 shared benchmarks.
DeepSeek
GoogleSummary
- They share 19 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 7 categories and Gemma 2 27B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 33.3 to 10.0.
- The biggest single-benchmark swing is BigCodeBench Instruct: 48.6% for DeepSeek-V2.5 (Sep 2024) and 42.8% for Gemma 2 27B.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 37.0 | 29.2 |
| Released | 2024-09-06 | 2024-06-24 |
| Weights | Open | Open |
| Context window | — | 8K |
| Max output | — | 2K |
| Input $ / M tokens | — | $0.65 |
| Output $ / M tokens | — | $0.65 |
| Results tracked | 22 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 2 27B leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), Gemma 2 27B: 33.5 (#243)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| BigCodeBench Instruct | 48.6% | 42.8% |
| LMArena Coding | 1310 | 1211 |
| BigCodeBench Complete | 53.2% | 52.5% |
| Aider Polyglot | 17.8% | — |
| LiveBench Coding | — | 36% |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), Gemma 2 27B: 14.8 (#313)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1199 |
| LiveBench Reasoning | — | 28.1% |
| DTBench | — | 48% |
| LiveBench Data Analysis | — | 47.9% |
| LMCA | — | 7.1% |
| Epoch Capabilities Index | — | 122.06 |
| LiveBench | — | 38.2% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), Gemma 2 27B: 10.0 (#306)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Math | 1289 | 1212 |
| OTIS Mock AIME 2024-2025 | — | 1.4% |
| LiveBench Math | — | 26.5% |
| MATH Level 5 | — | 27.9% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), Gemma 2 27B: 18.9 (#275)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Expert | 1266 | 1172 |
| GPQA Diamond | — | 36.5% |
| Confabulations | — | 27.1% |
| MMLU | — | 75.7% |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), Gemma 2 27B: 38.8 (#221)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Non-English | 1273 | 1217 |
| LMArena Chinese | 1317 | 1221 |
| LMArena French | 1288 | 1246 |
| LMArena German | 1258 | 1209 |
| LMArena Japanese | 1228 | 1175 |
| LMArena Korean | 1209 | 1174 |
| LMArena Russian | 1289 | 1234 |
| LMArena Spanish | 1248 | 1228 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), Gemma 2 27B: 60.7 (#244)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1206 |
| LiveBench Instruction Following | — | 58.1% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), Gemma 2 27B: 37.3 (#213)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Longer Query | 1301 | 1231 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), Gemma 2 27B: 44.4 (#220)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemma 2 27B |
|---|---|---|
| LMArena Text | 1294 | 1231 |
| LMArena Creative Writing | 1285 | 1241 |
| LMArena Multi-Turn | 1297 | 1224 |
| LiveBench Language | — | 32.6% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Gemma 2 27B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 29.2 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Gemma 2 27B better for coding?
Gemma 2 27B scores higher on coding benchmarks: 33.5 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Gemma 2 27B share?
19 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Gemma 2 27B has 34.