Model comparison
DeepSeek-V3.2-Exp vs Qwen3 Max
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 43.5 on the Noometry Index.
Last verified . 28 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 28 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 5 categories and Qwen3 Max in 3 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in long context, where DeepSeek-V3.2-Exp leads 47.6 to 41.6.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 52.2% for DeepSeek-V3.2-Exp and 72.5% for Qwen3 Max.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $1.20 / $6 for Qwen3 Max.
- Qwen3 Max accepts more context: 262K tokens versus 164K.
- DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Exp | Qwen3 Max | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 44.9 | 43.5 |
| Released | 2025-09-29 | 2025-09-23 |
| Weights | Open | Proprietary |
| Context window | 164K | 262K |
| Max output | 66K | 66K |
| Input $ / M tokens | $0.26 | $1.20 |
| Output $ / M tokens | $0.38 | $6 |
| Results tracked | 48 | 33 |
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), Qwen3 Max: 42.4 (#93)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| LMArena Coding | 1454 | 1456 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
| ALE-Bench | — | 370.45 |
Agentic & Tool Use Not comparable
DeepSeek-V3.2-Exp: 37.9 (#31), Qwen3 Max: —
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| Vending-Bench 2 | 1,034 | 71.56 |
| Terminal-Bench | 39.6% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
Reasoning Too close to call
DeepSeek-V3.2-Exp: 22.0 (#185), Qwen3 Max: 22.0 (#186)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 72.5% |
| NYT Connections (extended) | 36.7% | 30.1% |
| Chess Puzzles | 14% | 4% |
| LMArena Hard Prompts | 1433 | 1448 |
| DTBench | 87.7% | 82.1% |
| LMCA | 29.1% | 28.3% |
| Epoch Capabilities Index | 146.28 | 142.41 |
| ARC-AGI-2 | 4% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Thematic Generalization | 65% | — |
| Mystery Game Puzzles | — | 5% |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 40.8 (#75), Qwen3 Max: 37.6 (#94)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 73.3% |
| LMArena Math | 1435 | 1447 |
| FrontierMath (Tiers 1-3) | — | 18.9% |
| MathArena Final-Answer Competitions | 57.7% | — |
| ProofBench | 8% | — |
| MATH Level 5 | — | 97.1% |
| 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), Qwen3 Max: 48.1 (#72)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 83.4% | 72.6% |
| LMArena Expert | 1436 | 1457 |
| SimpleQA Verified | — | 48.7% |
| Vectara Hallucination Rate | 5.3% | — |
Multilingual Qwen3 Max leads
DeepSeek-V3.2-Exp: 52.4 (#85), Qwen3 Max: 53.9 (#58)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| LMArena Non-English | 1409 | 1429 |
| LMArena Chinese | 1461 | 1479 |
| LMArena French | 1435 | 1450 |
| LMArena German | 1440 | 1463 |
| LMArena Japanese | 1377 | 1399 |
| LMArena Korean | 1372 | 1397 |
| LMArena Russian | 1423 | 1428 |
| LMArena Spanish | 1441 | 1463 |
Instruction Following Too close to call
DeepSeek-V3.2-Exp: 74.7 (#89), Qwen3 Max: 75.0 (#81)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| LMArena Instruction Following | 1413 | 1420 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#15), Qwen3 Max: 41.6 (#129)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| Fiction.LiveBench | 83.3% | 66.7% |
| CL-bench | 13.2% | 14.5% |
| LMArena Longer Query | 1428 | 1439 |
| CL-bench Life | 9.5% | — |
Writing & Preference Too close to call
DeepSeek-V3.2-Exp: 62.5 (#73), Qwen3 Max: 62.6 (#72)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3 Max |
|---|---|---|
| LMArena Text | 1425 | 1439 |
| LMArena Creative Writing | 1403 | 1401 |
| LMArena Multi-Turn | 1427 | 1446 |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Qwen3 Max?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 43.5 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or Qwen3 Max?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Qwen3 Max lists at $1.20 and $6.
Is DeepSeek-V3.2-Exp or Qwen3 Max better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.4 versus 42.4 in the Noometry coding category.
Which has the bigger context window?
Qwen3 Max does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and Qwen3 Max share?
28 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 48 scored results on Noometry and Qwen3 Max has 33.