Model comparison
DeepSeek-V3.1 vs Qwen2.5-Max
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 40.7 on the Noometry Index.
Last verified . 18 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 6 categories and Qwen2.5-Max in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.1 leads 43.7 to 35.3.
- DeepSeek-V3.1 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.1 | Qwen2.5-Max | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.8 | 40.7 |
| Released | 2025-08-21 | 2025-01-25 |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.25 | — |
| Output $ / M tokens | $0.95 | — |
| Results tracked | 27 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5-Max leads
DeepSeek-V3.1: 40.3 (#144), Qwen2.5-Max: 41.8 (#117)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Coding | 1417 | 1359 |
| WeirdML | 38.4% | — |
| LiveBench Coding | — | 64.4% |
Reasoning DeepSeek-V3.1 leads
DeepSeek-V3.1: 27.9 (#110), Qwen2.5-Max: 25.6 (#147)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Hard Prompts | 1417 | 1360 |
| Epoch Capabilities Index | 139.92 | 132.53 |
| SimpleBench | 40% | — |
| Kagi LLM Benchmark | 53.2% | — |
| LiveBench Reasoning | — | 51.4% |
| DTBench | 82.7% | — |
| LiveBench Data Analysis | — | 67.9% |
| LMCA | 24.3% | — |
| ForecastBench | 58 | — |
| LiveBench | — | 62.3% |
Math DeepSeek-V3.1 leads
DeepSeek-V3.1: 38.9 (#122), Qwen2.5-Max: 36.9 (#162)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Math | 1420 | 1369 |
| LiveBench Math | — | 58.4% |
Knowledge DeepSeek-V3.1 leads
DeepSeek-V3.1: 43.7 (#90), Qwen2.5-Max: 35.3 (#186)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Expert | 1405 | 1337 |
| Confabulations | — | 21.8% |
| Vectara Hallucination Rate | 5.5% | — |
Multilingual DeepSeek-V3.1 leads
DeepSeek-V3.1: 51.6 (#106), Qwen2.5-Max: 48.1 (#146)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Non-English | 1400 | 1352 |
| LMArena Chinese | 1469 | 1382 |
| LMArena French | 1447 | 1396 |
| LMArena German | 1411 | 1350 |
| LMArena Japanese | 1378 | 1300 |
| LMArena Korean | 1337 | 1304 |
| LMArena Russian | 1405 | 1353 |
| LMArena Spanish | 1431 | 1377 |
Instruction Following DeepSeek-V3.1 leads
DeepSeek-V3.1: 73.9 (#110), Qwen2.5-Max: 71.3 (#152)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Instruction Following | 1400 | 1335 |
| LiveBench Instruction Following | — | 75.3% |
Long Context Qwen2.5-Max leads
DeepSeek-V3.1: 36.3 (#232), Qwen2.5-Max: 41.4 (#142)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Longer Query | 1422 | 1358 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference DeepSeek-V3.1 leads
DeepSeek-V3.1: 60.3 (#98), Qwen2.5-Max: 55.4 (#146)
| Benchmark | DeepSeek-V3.1 | Qwen2.5-Max |
|---|---|---|
| LMArena Text | 1420 | 1367 |
| LMArena Creative Writing | 1401 | 1339 |
| LMArena Multi-Turn | 1408 | 1364 |
| Short-Story Creative Writing | — | 72.9% |
| EQ-Bench Creative Writing | 1436 | — |
| LiveBench Language | — | 56.3% |
Frequently asked questions
Is DeepSeek-V3.1 better than Qwen2.5-Max?
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 40.7 on the Noometry Index.
Is DeepSeek-V3.1 or Qwen2.5-Max better for coding?
Qwen2.5-Max scores higher on coding benchmarks: 41.8 versus 40.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.1 and Qwen2.5-Max share?
18 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Qwen2.5-Max has 27.