Model comparison
DeepSeek-V3.2-Exp vs QwQ-32B
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 39.4 on the Noometry Index.
Last verified . 24 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 6 categories and QwQ-32B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.6 to 37.2.
- The biggest single-benchmark swing is Aider Polyglot: 74.2% for DeepSeek-V3.2-Exp and 20.9% for QwQ-32B.
Side by side
| DeepSeek-V3.2-Exp | QwQ-32B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 44.9 | 39.4 |
| Released | 2025-09-29 | 2024-11-28 |
| Weights | Open | Open |
| Context window | 164K | — |
| Max output | 66K | — |
| Input $ / M tokens | $0.26 | — |
| Output $ / M tokens | $0.38 | — |
| Results tracked | 48 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.4 (#61), QwQ-32B: 34.9 (#220)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| Aider Polyglot | 74.2% | 20.9% |
| LMArena Coding | 1454 | 1333 |
| SWE-bench Verified (bash only) | 70% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
| BigCodeBench Instruct | — | 44.6% |
| LiveBench Coding | — | 72.2% |
| BigCodeBench Complete | — | 54.4% |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 37.9 (#31), QwQ-32B: —
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Too close to call
DeepSeek-V3.2-Exp: 22.0 (#185), QwQ-32B: 22.9 (#171)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| Chess Puzzles | 14% | 5% |
| LMArena Hard Prompts | 1433 | 1326 |
| Epoch Capabilities Index | 146.28 | 137.6 |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Thematic Generalization | 65% | — |
| LiveBench Reasoning | — | 83.5% |
| DTBench | 87.7% | — |
| LiveBench Data Analysis | — | 65% |
| LMCA | 29.1% | — |
| ForecastBench | — | 58.3 |
| LiveBench | — | 72% |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 40.8 (#75), QwQ-32B: 36.2 (#123)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 59.2% |
| LMArena Math | 1435 | 1359 |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| LiveBench Math | — | 77.8% |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.6 (#61), QwQ-32B: 37.2 (#150)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| GPQA Diamond | 83.4% | 65.3% |
| LMArena Expert | 1436 | 1323 |
| Confabulations | — | 15.6% |
| Vectara Hallucination Rate | 5.3% | — |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.4 (#85), QwQ-32B: 45.0 (#171)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1409 | 1306 |
| LMArena Chinese | 1461 | 1378 |
| LMArena French | 1435 | 1337 |
| LMArena German | 1440 | 1313 |
| LMArena Japanese | 1377 | 1263 |
| LMArena Korean | 1372 | 1278 |
| LMArena Russian | 1423 | 1297 |
| LMArena Spanish | 1441 | 1353 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.7 (#89), QwQ-32B: 72.8 (#131)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| LMArena Instruction Following | 1413 | 1297 |
| LiveBench Instruction Following | — | 81.8% |
Long Context QwQ-32B leads
DeepSeek-V3.2-Exp: 47.6 (#15), QwQ-32B: 49.0 (#11)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| Fiction.LiveBench | 83.3% | 83.3% |
| LMArena Longer Query | 1428 | 1308 |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.5 (#73), QwQ-32B: 50.9 (#175)
| Benchmark | DeepSeek-V3.2-Exp | QwQ-32B |
|---|---|---|
| LMArena Text | 1425 | 1329 |
| LMArena Creative Writing | 1403 | 1288 |
| EQ-Bench Creative Writing | 1515 | 1257 |
| LMArena Multi-Turn | 1427 | 1314 |
| Short-Story Creative Writing | — | 80.2% |
| LiveBench Language | — | 51.4% |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than QwQ-32B?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 39.4 on the Noometry Index.
Is DeepSeek-V3.2-Exp or QwQ-32B better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.4 versus 34.9 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Exp and QwQ-32B share?
24 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 48 scored results on Noometry and QwQ-32B has 36.