Model comparison
DeepSeek-V2.5 (Sep 2024) vs QwQ-32B
QwQ-32B is the stronger model overall, scoring 39.8 to 37.6 on the Noometry Index.
Last verified . 20 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 20 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 1 category and QwQ-32B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in long context, where QwQ-32B leads 49.0 to 39.5.
Side by side
| DeepSeek-V2.5 (Sep 2024) | QwQ-32B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 37.6 | 39.8 |
| Released | 2024-09-06 | 2024-11-28 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), QwQ-32B: 35.4 (#226)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| Aider Polyglot | 17.8% | 20.9% |
| BigCodeBench Instruct | 48.6% | 44.6% |
| LMArena Coding | 1309 | 1333 |
| BigCodeBench Complete | 53.2% | 54.4% |
| LiveBench Coding | — | 72.2% |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), QwQ-32B: 23.7 (#174)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1325 |
| Chess Puzzles | — | 5% |
| LiveBench Reasoning | — | 83.5% |
| LiveBench Data Analysis | — | 65% |
| Epoch Capabilities Index | — | 137.6 |
| ForecastBench | — | 58.3 |
| LiveBench | — | 72% |
Math QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), QwQ-32B: 38.0 (#143)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Math | 1288 | 1359 |
| OTIS Mock AIME 2024-2025 | — | 59.2% |
| LiveBench Math | — | 77.8% |
Knowledge QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), QwQ-32B: 37.2 (#158)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Expert | 1266 | 1324 |
| GPQA Diamond | — | 65.3% |
| Confabulations | — | 15.6% |
Multilingual QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), QwQ-32B: 44.8 (#176)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1273 | 1305 |
| LMArena Chinese | 1318 | 1378 |
| LMArena French | 1289 | 1336 |
| LMArena German | 1258 | 1313 |
| LMArena Japanese | 1228 | 1262 |
| LMArena Korean | 1209 | 1279 |
| LMArena Russian | 1289 | 1297 |
| LMArena Spanish | 1248 | 1354 |
Instruction Following QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), QwQ-32B: 72.6 (#137)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1297 |
| LiveBench Instruction Following | — | 81.8% |
Long Context QwQ-32B leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), QwQ-32B: 49.0 (#11)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Longer Query | 1301 | 1308 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference Too close to call
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), QwQ-32B: 50.6 (#180)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | QwQ-32B |
|---|---|---|
| LMArena Text | 1294 | 1329 |
| LMArena Creative Writing | 1285 | 1288 |
| LMArena Multi-Turn | 1297 | 1314 |
| Short-Story Creative Writing | — | 80.2% |
| EQ-Bench Creative Writing | — | 1257 |
| LiveBench Language | — | 51.4% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than QwQ-32B?
QwQ-32B is the stronger model overall, scoring 39.8 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or QwQ-32B better for coding?
QwQ-32B scores higher on coding benchmarks: 35.4 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and QwQ-32B share?
20 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and QwQ-32B has 36.