Model comparison
DeepSeek-V2.5 (Sep 2024) vs Qwen Max
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 34.7 on the Noometry Index.
Last verified . 18 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 8 categories and Qwen Max in 0 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 35.9 to 22.3.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Qwen Max | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 37.6 | 34.7 |
| Released | 2024-09-06 | 2024-04-03 |
| Weights | Open | Proprietary |
| Context window | — | 33K |
| Max output | — | 8K |
| Input $ / M tokens | — | $1.60 |
| Output $ / M tokens | — | $6.40 |
| Results tracked | 22 | 23 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Qwen Max: 30.7 (#292)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| Aider Polyglot | 17.8% | 21.8% |
| LMArena Coding | 1309 | 1288 |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning Too close to call
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Qwen Max: 25.1 (#151)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1269 |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Qwen Max: 22.3 (#276)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Math | 1288 | 1275 |
| OTIS Mock AIME 2024-2025 | — | 16.1% |
| MATH Level 5 | — | 67.2% |
| FrontierMath (Feb 2025 set) | — | 1% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Qwen Max: 30.3 (#228)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Expert | 1266 | 1248 |
| GPQA Diamond | — | 56.1% |
Multilingual Too close to call
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Qwen Max: 41.8 (#202)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Non-English | 1273 | 1263 |
| LMArena Chinese | 1318 | 1254 |
| LMArena French | 1289 | 1330 |
| LMArena German | 1258 | 1254 |
| LMArena Japanese | 1228 | 1205 |
| LMArena Korean | 1209 | 1142 |
| LMArena Russian | 1289 | 1274 |
| LMArena Spanish | 1248 | 1290 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Qwen Max: 66.5 (#208)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Instruction Following | 1280 | 1262 |
Long Context Too close to call
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Qwen Max: 39.4 (#180)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Longer Query | 1301 | 1288 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Qwen Max: 47.8 (#205)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen Max |
|---|---|---|
| LMArena Text | 1294 | 1282 |
| LMArena Creative Writing | 1285 | 1248 |
| LMArena Multi-Turn | 1297 | 1277 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Qwen Max?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.6 to 34.7 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Qwen Max better for coding?
They score almost the same on coding (31.7 vs 30.7); test both on your own repository before choosing.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Qwen Max share?
18 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Qwen Max has 23.