Model comparison
DeepSeek-V3 vs Qwen3 32B
DeepSeek-V3 and Qwen3 32B score almost the same on the Noometry Index (39.5 vs 39.2), so choose on price, context window or the category you care about most.
Last verified . 24 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. DeepSeek-V3 scores higher in 5 categories and Qwen3 32B in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in long context, where Qwen3 32B leads 43.8 to 34.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 66.9% for Qwen3 32B.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.70 / $2.80 for Qwen3 32B.
- DeepSeek-V3 accepts more context: 164K tokens versus 131K.
Side by side
| DeepSeek-V3 | Qwen3 32B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 39.5 | 39.2 |
| Released | 2024-12-26 | 2025-04 |
| Weights | Open | Open |
| Context window | 164K | 131K |
| Max output | 164K | 16K |
| Input $ / M tokens | $0.24 | $0.70 |
| Output $ / M tokens | $0.90 | $2.80 |
| Results tracked | 60 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3 leads
DeepSeek-V3: 42.3 (#106), Qwen3 32B: 37.7 (#190)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| Aider Polyglot | 55.1% | 40% |
| SciCode | 35.8% | 35.4% |
| LMArena Coding | 1368 | 1358 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, Qwen3 32B: 32.6 (#62)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |
| METR Time Horizons | 49.6% | — |
Reasoning Too close to call
DeepSeek-V3: 20.5 (#236), Qwen3 32B: 20.2 (#241)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | 52.3% | 54.9% |
| CritPt | 0% | 0.3% |
| LMArena Hard Prompts | 1365 | 1334 |
| DTBench | 64.8% | 67.5% |
| LMCA | 15.5% | 17.3% |
| Epoch Capabilities Index | 135.94 | 138.51 |
| SimpleBench | 27.2% | — |
| Chess Puzzles | — | 5% |
| LiveBench Reasoning | 65.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math Qwen3 32B leads
DeepSeek-V3: 32.1 (#219), Qwen3 32B: 39.7 (#99)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 66.9% |
| LMArena Math | 1373 | 1399 |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge Qwen3 32B leads
DeepSeek-V3: 37.5 (#155), Qwen3 32B: 40.0 (#125)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 67.6% | 65.7% |
| Vectara Hallucination Rate | 6.1% | 5.9% |
| LMArena Expert | 1351 | 1362 |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multilingual DeepSeek-V3 leads
DeepSeek-V3: 48.5 (#143), Qwen3 32B: 45.6 (#167)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1358 | 1317 |
| LMArena Chinese | 1391 | 1357 |
| LMArena German | 1374 | 1341 |
| LMArena Russian | 1373 | 1311 |
| LMArena French | 1385 | — |
| LMArena Japanese | 1333 | — |
| LMArena Korean | 1319 | — |
| LMArena Spanish | 1358 | — |
Instruction Following DeepSeek-V3 leads
DeepSeek-V3: 72.8 (#130), Qwen3 32B: 68.9 (#179)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1345 | 1305 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context Qwen3 32B leads
DeepSeek-V3: 34.0 (#253), Qwen3 32B: 43.8 (#87)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | 50% | 74.2% |
| LMArena Longer Query | 1352 | 1327 |
Writing & Preference DeepSeek-V3 leads
DeepSeek-V3: 57.4 (#130), Qwen3 32B: 52.9 (#163)
| Benchmark | DeepSeek-V3 | Qwen3 32B |
|---|---|---|
| LMArena Text | 1375 | 1340 |
| LMArena Creative Writing | 1364 | 1297 |
| LMArena Multi-Turn | 1389 | 1331 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than Qwen3 32B?
DeepSeek-V3 and Qwen3 32B score almost the same on the Noometry Index (39.5 vs 39.2), so choose on price, context window or the category you care about most.
Which is cheaper, DeepSeek-V3 or Qwen3 32B?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.
Is DeepSeek-V3 or Qwen3 32B better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 37.7 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3 does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-V3 and Qwen3 32B share?
24 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Qwen3 32B has 26.