Model comparison
DeepSeek V4 Pro vs Qwen2.5-Coder-32B
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 33.4 on the Noometry Index.
Last verified . 13 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and Qwen2.5-Coder-32B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 21.2.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 33K.
Side by side
| DeepSeek V4 Pro | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 54.3 | 33.4 |
| Released | 2026-04-24 | 2024-09-18 |
| Weights | Open | Open |
| Context window | 1M | 33K |
| Max output | 393K | 29K |
| Input $ / M tokens | $0.66 | $0.66 |
| Output $ / M tokens | $1.98 | $1 |
| Results tracked | 48 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Coding | 1470 | 1276 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| WeirdML | 66.2% | — |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| ALE-Bench | 1,403 | — |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), Qwen2.5-Coder-32B: —
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1461 | 1251 |
| Epoch Capabilities Index | 155.31 | 119.49 |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| CritPt | 18% | — |
| Chess Puzzles | 47% | — |
| LiveBench Reasoning | — | 42.1% |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LiveBench Data Analysis | — | 49.9% |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1455 | 1251 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
| LiveBench Math | — | 46.6% |
| GSM8K | — | 93% |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1464 | 1221 |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1439 | 1205 |
| LMArena Chinese | 1486 | 1222 |
| LMArena Russian | 1453 | 1228 |
| LMArena French | 1472 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1445 | — |
| LMArena Korean | 1447 | — |
| LMArena Spanish | 1458 | — |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1448 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1458 | 1251 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | DeepSeek V4 Pro | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1451 | 1230 |
| LMArena Creative Writing | 1446 | 1174 |
| LMArena Multi-Turn | 1467 | 1222 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is DeepSeek V4 Pro better than Qwen2.5-Coder-32B?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 33.4 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; DeepSeek V4 Pro lists at $0.66 and $1.98.
Is DeepSeek V4 Pro or Qwen2.5-Coder-32B better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 33K.
How many benchmarks do DeepSeek V4 Pro and Qwen2.5-Coder-32B share?
13 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Qwen2.5-Coder-32B has 31.