Model comparison
Deepseek Coder v2 vs Qwen1.5-110B
Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 33.6 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 8 categories and Qwen1.5-110B in 0 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in coding, where Deepseek Coder v2 leads 37.5 to 32.4.
- The biggest single-benchmark swing is BigCodeBench Complete: 59.7% for Deepseek Coder v2 and 44.4% for Qwen1.5-110B.
Side by side
| Deepseek Coder v2 | Qwen1.5-110B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 35.3 | 33.6 |
| Released | 2024-06-17 | 2024-04-25 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 20 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Deepseek Coder v2 leads
Deepseek Coder v2: 37.5 (#178), Qwen1.5-110B: 32.4 (#260)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| BigCodeBench Instruct | 48.2% | 35% |
| LMArena Coding | 1251 | 1184 |
| BigCodeBench Complete | 59.7% | 44.4% |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Too close to call
Deepseek Coder v2: 22.6 (#175), Qwen1.5-110B: 21.7 (#192)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1168 |
| ForecastBench | — | 57.7 |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#188), Qwen1.5-110B: 31.1 (#199)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Math | 1241 | 1185 |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 31.7 (#209), Qwen1.5-110B: 30.7 (#217)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Expert | 1181 | 1144 |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual Deepseek Coder v2 leads
Deepseek Coder v2: 36.4 (#235), Qwen1.5-110B: 33.8 (#245)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Non-English | 1182 | 1142 |
| LMArena Chinese | 1201 | 1206 |
| LMArena French | 1184 | 1150 |
| LMArena German | 1164 | 1123 |
| LMArena Japanese | 1126 | 1074 |
| LMArena Korean | 1104 | 1044 |
| LMArena Russian | 1188 | 1118 |
| LMArena Spanish | 1153 | 1141 |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.9 (#237), Qwen1.5-110B: 60.6 (#247)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1158 |
Long Context Deepseek Coder v2 leads
Deepseek Coder v2: 37.0 (#219), Qwen1.5-110B: 35.0 (#237)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Longer Query | 1219 | 1157 |
Writing & Preference Too close to call
Deepseek Coder v2: 38.4 (#249), Qwen1.5-110B: 38.1 (#250)
| Benchmark | Deepseek Coder v2 | Qwen1.5-110B |
|---|---|---|
| LMArena Text | 1191 | 1175 |
| LMArena Creative Writing | 1120 | 1148 |
| LMArena Multi-Turn | 1176 | 1160 |
Frequently asked questions
Is Deepseek Coder v2 better than Qwen1.5-110B?
Deepseek Coder v2 is the stronger model overall, scoring 35.3 to 33.6 on the Noometry Index.
Is Deepseek Coder v2 or Qwen1.5-110B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 37.5 versus 32.4 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Qwen1.5-110B share?
19 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Qwen1.5-110B has 20.