Model comparison
DeepSeek-V3.2-Exp vs Qwen Max
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 34.7 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 7 categories and Qwen Max in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 30.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 87.8% for DeepSeek-V3.2-Exp and 16.1% for Qwen Max.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $1.60 / $6.40 for Qwen Max.
- DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 33K.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Exp | Qwen Max | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 44.3 | 34.7 |
| Released | 2025-09-29 | 2024-04-03 |
| Weights | Open | Proprietary |
| Context window | 164K | 33K |
| Max output | 66K | 8K |
| Input $ / M tokens | $0.26 | $1.60 |
| Output $ / M tokens | $0.38 | $6.40 |
| Results tracked | 49 | 23 |
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: 46.5 (#65), Qwen Max: 30.7 (#292)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| Aider Polyglot | 74.2% | 21.8% |
| LMArena Coding | 1454 | 1288 |
| SWE-bench Verified (bash only) | 70% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 32.7 (#59), Qwen Max: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Qwen Max leads
DeepSeek-V3.2-Exp: 22.1 (#208), Qwen Max: 25.1 (#151)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| LMArena Hard Prompts | 1434 | 1269 |
| ARC-AGI-2 | 4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Chess Puzzles | 14% | — |
| Thematic Generalization | 65% | — |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
| Epoch Capabilities Index | 146.27 | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), Qwen Max: 22.3 (#276)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 16.1% |
| LMArena Math | 1435 | 1275 |
| FrontierMath (Feb 2025 set) | 22.1% | 1% |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| MATH Level 5 | — | 67.2% |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), Qwen Max: 30.3 (#228)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| GPQA Diamond | 83.4% | 56.1% |
| LMArena Expert | 1436 | 1248 |
| Vectara Hallucination Rate | 5.3% | — |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), Qwen Max: 41.8 (#202)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| LMArena Non-English | 1409 | 1263 |
| LMArena Chinese | 1461 | 1254 |
| LMArena French | 1433 | 1330 |
| LMArena German | 1440 | 1254 |
| LMArena Japanese | 1374 | 1205 |
| LMArena Korean | 1371 | 1142 |
| LMArena Russian | 1424 | 1274 |
| LMArena Spanish | 1440 | 1290 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), Qwen Max: 66.5 (#208)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| LMArena Instruction Following | 1413 | 1262 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), Qwen Max: 39.4 (#180)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| Fiction.LiveBench | 83.3% | 66.7% |
| LMArena Longer Query | 1428 | 1288 |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Qwen Max: 47.8 (#205)
| Benchmark | DeepSeek-V3.2-Exp | Qwen Max |
|---|---|---|
| LMArena Text | 1425 | 1282 |
| LMArena Creative Writing | 1403 | 1248 |
| LMArena Multi-Turn | 1427 | 1277 |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Qwen Max?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 34.7 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or Qwen Max?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Qwen Max lists at $1.60 and $6.40.
Is DeepSeek-V3.2-Exp or Qwen Max better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 30.7 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3.2-Exp does, with 164K tokens against 33K.
How many benchmarks do DeepSeek-V3.2-Exp and Qwen Max share?
22 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Qwen Max has 23.