Model comparison
Deepseek Coder v2 vs Qwen2.5-Coder-32B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 33.4 on the Noometry Index.
Last verified . 19 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Deepseek Coder v2 scores higher in 4 categories and Qwen2.5-Coder-32B in 4 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in coding, where Deepseek Coder v2 leads 38.1 to 22.6.
Side by side
| Deepseek Coder v2 | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 35.9 | 33.4 |
| Released | 2024-06-17 | 2024-09-18 |
| Weights | Open | Open |
| Context window | — | 33K |
| Max output | — | 29K |
| Input $ / M tokens | — | $0.66 |
| Output $ / M tokens | — | $1 |
| Results tracked | 24 | 31 |
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), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| BigCodeBench Instruct | 48.2% | 49% |
| LMArena Coding | 1251 | 1276 |
| BigCodeBench Complete | 59.7% | 58% |
| HumanEval+ | 82.3% | 87.2% |
| MBPP+ | 75.1% | 77% |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| LiveBench Coding | — | 56.9% |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1251 |
| WinoGrande | 83.7% | 80.8% |
| LiveBench Reasoning | — | 42.1% |
| LiveBench Data Analysis | — | 49.9% |
| Epoch Capabilities Index | — | 119.49 |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1241 | 1251 |
| GSM8K | 94.5% | 93% |
| LiveBench Math | — | 46.6% |
Knowledge Qwen2.5-Coder-32B leads
Deepseek Coder v2: 32.3 (#212), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1181 | 1221 |
| ARC (AI2) Challenge | 64.3% | 70.5% |
| MMLU | — | 79.1% |
Multilingual Qwen2.5-Coder-32B leads
Deepseek Coder v2: 36.3 (#240), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1182 | 1205 |
| LMArena Chinese | 1201 | 1222 |
| LMArena Russian | 1188 | 1228 |
| LMArena French | 1185 | — |
| LMArena German | 1164 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Too close to call
Deepseek Coder v2: 61.7 (#242), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
Long Context Too close to call
Deepseek Coder v2: 37.0 (#224), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1219 | 1251 |
Writing & Preference Qwen2.5-Coder-32B leads
Deepseek Coder v2: 38.2 (#253), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | Deepseek Coder v2 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1191 | 1230 |
| LMArena Creative Writing | 1120 | 1174 |
| LMArena Multi-Turn | 1177 | 1222 |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Deepseek Coder v2 better than Qwen2.5-Coder-32B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 33.4 on the Noometry Index.
Is Deepseek Coder v2 or Qwen2.5-Coder-32B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 22.6 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Qwen2.5-Coder-32B share?
19 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Qwen2.5-Coder-32B has 31.