Model comparison
Deepseek Coder v2 vs Qwen2.5 7B Instruct
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.0 on the Noometry Index.
Last verified . 2 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. Deepseek Coder v2 scores higher in 4 categories and Qwen2.5 7B Instruct in 2 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Deepseek Coder v2 leads 34.9 to 12.6.
- The biggest single-benchmark swing is BigCodeBench Complete: 59.7% for Deepseek Coder v2 and 46.1% for Qwen2.5 7B Instruct.
Side by side
| Deepseek Coder v2 | Qwen2.5 7B Instruct | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 35.9 | 29.0 |
| Released | 2024-06-17 | 2024-09 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $0.17 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 24 | 15 |
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 7B Instruct: 36.5 (#208)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| BigCodeBench Instruct | 48.2% | 37.6% |
| BigCodeBench Complete | 59.7% | 46.1% |
| LMArena Coding | 1251 | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, Qwen2.5 7B Instruct: 23.8 (#124)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| BALROG | — | 7.8% |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Qwen2.5 7B Instruct: 14.8 (#322)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| Chess Puzzles | — | 0% |
| LMArena Hard Prompts | 1207 | — |
| DTBench | — | 47.7% |
| LMCA | — | 6.4% |
| Epoch Capabilities Index | — | 118.51 |
| WinoGrande | 83.7% | — |
Math Deepseek Coder v2 leads
Deepseek Coder v2: 34.9 (#190), Qwen2.5 7B Instruct: 12.6 (#306)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 2.5% |
| Omni-MATH | — | 29.4% |
| LMArena Math | 1241 | — |
| GSM8K | 94.5% | — |
Knowledge Deepseek Coder v2 leads
Deepseek Coder v2: 32.3 (#212), Qwen2.5 7B Instruct: 17.0 (#286)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| GPQA Diamond | — | 35.5% |
| MMLU-Pro | — | 53.9% |
| GPQA (HELM) | — | 34.1% |
| LMArena Expert | 1181 | — |
| ARC (AI2) Challenge | 64.3% | — |
| MMLU | — | 72.9% |
Multilingual Not comparable
Deepseek Coder v2: 36.3 (#240), Qwen2.5 7B Instruct: —
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Non-English | 1182 | — |
| LMArena Chinese | 1201 | — |
| LMArena French | 1185 | — |
| LMArena German | 1164 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Russian | 1188 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Qwen2.5 7B Instruct leads
Deepseek Coder v2: 61.7 (#242), Qwen2.5 7B Instruct: 63.2 (#231)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| IFEval | — | 74.1% |
| LMArena Instruction Following | 1180 | — |
Long Context Not comparable
Deepseek Coder v2: 37.0 (#224), Qwen2.5 7B Instruct: —
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Longer Query | 1219 | — |
Writing & Preference Qwen2.5 7B Instruct leads
Deepseek Coder v2: 38.2 (#253), Qwen2.5 7B Instruct: 48.8 (#195)
| Benchmark | Deepseek Coder v2 | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Text | 1191 | — |
| LMArena Creative Writing | 1120 | — |
| WildBench | — | 73.1% |
| LMArena Multi-Turn | 1177 | — |
Frequently asked questions
Is Deepseek Coder v2 better than Qwen2.5 7B Instruct?
Deepseek Coder v2 is the stronger model overall, scoring 35.9 to 29.0 on the Noometry Index.
Is Deepseek Coder v2 or Qwen2.5 7B Instruct better for coding?
Deepseek Coder v2 scores higher on coding benchmarks: 38.1 versus 36.5 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Qwen2.5 7B Instruct share?
2 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Qwen2.5 7B Instruct has 15.