Model comparison
Deepseek Coder v2 vs Gemma 3 4B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 28.1 on the Noometry Index.
Last verified . 12 shared benchmarks.
DeepSeek
GoogleSummary
- They share 12 benchmarks with published results for both. Deepseek Coder v2 scores higher in 4 categories and Gemma 3 4B in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Deepseek Coder v2 leads 32.3 to 11.8.
Side by side
| Deepseek Coder v2 | Gemma 3 4B | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 35.9 | 28.1 |
| Released | 2024-06-17 | 2025-03-12 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.04 |
| Output $ / M tokens | — | $0.08 |
| Results tracked | 24 | 22 |
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 3 4B: 35.9 (#215)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1251 | 1230 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, Gemma 3 4B: 20.9 (#142)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 19.6% |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Gemma 3 4B: 13.2 (#335)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1253 |
| Kagi LLM Benchmark | — | 25.2% |
| Chess Puzzles | — | 0% |
| DTBench | — | 50.9% |
| LMCA | — | 2.8% |
| Epoch Capabilities Index | — | 116.02 |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Gemma 3 4B: 16.8 (#292)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Math | 1241 | 1239 |
| OTIS Mock AIME 2024-2025 | — | 7.5% |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Gemma 3 4B: 11.8 (#299)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Expert | 1181 | 1223 |
| GPQA Diamond | — | 23.2% |
| Vectara Hallucination Rate | — | 6.4% |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual Gemma 3 4B leads
Deepseek Coder v2: 36.3 (#240), Gemma 3 4B: 42.5 (#194)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1182 | 1273 |
| LMArena German | 1164 | 1281 |
| LMArena Russian | 1188 | 1294 |
| LMArena Chinese | 1201 | — |
| LMArena French | 1185 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Gemma 3 4B leads
Deepseek Coder v2: 61.7 (#242), Gemma 3 4B: 65.2 (#225)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1239 |
Long Context Gemma 3 4B leads
Deepseek Coder v2: 37.0 (#224), Gemma 3 4B: 38.7 (#194)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1219 | 1273 |
Writing & Preference Gemma 3 4B leads
Deepseek Coder v2: 38.2 (#253), Gemma 3 4B: 42.0 (#239)
| Benchmark | Deepseek Coder v2 | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1191 | 1291 |
| LMArena Creative Writing | 1120 | 1271 |
| LMArena Multi-Turn | 1177 | 1255 |
| EQ-Bench Creative Writing | — | 1068 |
Frequently asked questions
Is Deepseek Coder v2 better than Gemma 3 4B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 28.1 on the Noometry Index.
Is Deepseek Coder v2 or Gemma 3 4B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 35.9 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Gemma 3 4B share?
12 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Gemma 3 4B has 22.