Model comparison
Deepseek Coder v2 vs Qwen3.5 397B-A17B
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 35.9 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. Deepseek Coder v2 scores higher in 0 categories and Qwen3.5 397B-A17B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.5 397B-A17B leads 62.3 to 38.2.
Side by side
| Deepseek Coder v2 | Qwen3.5 397B-A17B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 35.9 | 46.0 |
| Released | 2024-06-17 | 2026-02-01 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.60 |
| Output $ / M tokens | — | $3.60 |
| Results tracked | 24 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 397B-A17B leads
Deepseek Coder v2: 38.1 (#183), Qwen3.5 397B-A17B: 42.0 (#114)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Coding | 1251 | 1465 |
| LMArena WebDev | — | 1400 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, Qwen3.5 397B-A17B: 33.3 (#53)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| APEX-Agents | — | 24.9% |
| τ²-bench Airline | — | 81.5% |
| τ²-bench Banking | — | 9.8% |
| τ²-bench Retail | — | 84.4% |
| τ²-bench Telecom | — | 97.8% |
Reasoning Qwen3.5 397B-A17B leads
Deepseek Coder v2: 23.6 (#176), Qwen3.5 397B-A17B: 34.5 (#70)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1448 |
| Kagi LLM Benchmark | — | 73.7% |
| NYT Connections (extended) | — | 58.9% |
| Chess Puzzles | — | 13% |
| Thematic Generalization | — | 65.1% |
| Mystery Game Puzzles | — | 18% |
| DTBench | — | 87.5% |
| LMCA | — | 37.9% |
| Epoch Capabilities Index | — | 146.65 |
| WinoGrande | 83.7% | — |
Math Qwen3.5 397B-A17B leads
Deepseek Coder v2: 34.9 (#190), Qwen3.5 397B-A17B: 46.1 (#73)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Math | 1241 | 1454 |
| FrontierMath (Tiers 1-3) | — | 31.2% |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| GSM8K | 94.5% | — |
Knowledge Qwen3.5 397B-A17B leads
Deepseek Coder v2: 32.3 (#212), Qwen3.5 397B-A17B: 53.3 (#58)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Expert | 1181 | 1462 |
| GPQA Diamond | — | 86.4% |
| ARC (AI2) Challenge | 64.3% | — |
Multimodal Not comparable
Deepseek Coder v2: —, Qwen3.5 397B-A17B: 40.7 (#44)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Vision | — | 1263 |
Multilingual Qwen3.5 397B-A17B leads
Deepseek Coder v2: 36.3 (#240), Qwen3.5 397B-A17B: 53.7 (#59)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Non-English | 1182 | 1430 |
| LMArena Chinese | 1201 | 1500 |
| LMArena French | 1185 | 1461 |
| LMArena German | 1164 | 1447 |
| LMArena Japanese | 1126 | 1426 |
| LMArena Korean | 1104 | 1384 |
| LMArena Russian | 1188 | 1429 |
| LMArena Spanish | 1153 | 1441 |
Instruction Following Qwen3.5 397B-A17B leads
Deepseek Coder v2: 61.7 (#242), Qwen3.5 397B-A17B: 75.0 (#77)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Instruction Following | 1180 | 1424 |
Long Context Qwen3.5 397B-A17B leads
Deepseek Coder v2: 37.0 (#224), Qwen3.5 397B-A17B: 44.1 (#74)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Longer Query | 1219 | 1442 |
Writing & Preference Qwen3.5 397B-A17B leads
Deepseek Coder v2: 38.2 (#253), Qwen3.5 397B-A17B: 62.3 (#79)
| Benchmark | Deepseek Coder v2 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Text | 1191 | 1438 |
| LMArena Creative Writing | 1120 | 1401 |
| LMArena Multi-Turn | 1177 | 1446 |
| EQ-Bench Creative Writing | — | 1478 |
Frequently asked questions
Is Deepseek Coder v2 better than Qwen3.5 397B-A17B?
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 35.9 on the Noometry Index.
Is Deepseek Coder v2 or Qwen3.5 397B-A17B better for coding?
Qwen3.5 397B-A17B scores higher on coding benchmarks: 42.0 versus 38.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Qwen3.5 397B-A17B share?
17 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Qwen3.5 397B-A17B has 36.