Model comparison
Deepseek Coder v2 vs Qwen1.5-14B
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 32.7 on the Noometry Index.
Last verified . 16 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 16 benchmarks with published results for both. Deepseek Coder v2 scores higher in 8 categories and Qwen1.5-14B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Deepseek Coder v2 leads 36.3 to 30.7.
Side by side
| Deepseek Coder v2 | Qwen1.5-14B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 35.9 | 32.7 |
| Released | 2024-06-17 | 2024-02-04 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 17 |
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), Qwen1.5-14B: 33.1 (#263)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Coding | 1251 | 1138 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Qwen1.5-14B: 21.4 (#223)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1113 |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Qwen1.5-14B: 32.4 (#215)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Math | 1241 | 1125 |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Qwen1.5-14B: 29.8 (#232)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Expert | 1181 | 1094 |
| ARC (AI2) Challenge | 64.3% | — |
| MMLU | — | 68.6% |
Multilingual Deepseek Coder v2 leads
Deepseek Coder v2: 36.3 (#240), Qwen1.5-14B: 30.7 (#262)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Non-English | 1182 | 1095 |
| LMArena Chinese | 1201 | 1147 |
| LMArena French | 1185 | 1116 |
| LMArena German | 1164 | 1043 |
| LMArena Japanese | 1126 | 1019 |
| LMArena Russian | 1188 | 1046 |
| LMArena Spanish | 1153 | 1085 |
| LMArena Korean | 1104 | — |
Instruction Following Deepseek Coder v2 leads
Deepseek Coder v2: 61.7 (#242), Qwen1.5-14B: 56.8 (#271)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1102 |
Long Context Deepseek Coder v2 leads
Deepseek Coder v2: 37.0 (#224), Qwen1.5-14B: 33.7 (#257)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Longer Query | 1219 | 1113 |
Writing & Preference Deepseek Coder v2 leads
Deepseek Coder v2: 38.2 (#253), Qwen1.5-14B: 33.6 (#276)
| Benchmark | Deepseek Coder v2 | Qwen1.5-14B |
|---|---|---|
| LMArena Text | 1191 | 1128 |
| LMArena Creative Writing | 1120 | 1091 |
| LMArena Multi-Turn | 1177 | 1110 |
Frequently asked questions
Is Deepseek Coder v2 better than Qwen1.5-14B?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 32.7 on the Noometry Index.
Is Deepseek Coder v2 or Qwen1.5-14B better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 33.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Qwen1.5-14B share?
16 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Qwen1.5-14B has 17.