Model comparison
DeepSeek-V3.2-Exp vs Qwen3-30B-A3B
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 38.9 on the Noometry Index.
Last verified . 30 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 30 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 8 categories and Qwen3-30B-A3B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in long context, where DeepSeek-V3.2-Exp leads 47.6 to 31.0.
- The biggest single-benchmark swing is Fiction.LiveBench: 83.3% for DeepSeek-V3.2-Exp and 40.6% for Qwen3-30B-A3B.
- Qwen3-30B-A3B is cheaper at $0.12 / $0.50 per million input/output tokens, against $0.26 / $0.38 for DeepSeek-V3.2-Exp.
- DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 41K.
Side by side
| DeepSeek-V3.2-Exp | Qwen3-30B-A3B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 44.3 | 38.9 |
| Released | 2025-09-29 | 2025-04-28 |
| Weights | Open | Open |
| Context window | 164K | 41K |
| Max output | 66K | 16K |
| Input $ / M tokens | $0.26 | $0.12 |
| Output $ / M tokens | $0.38 | $0.50 |
| Results tracked | 49 | 32 |
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), Qwen3-30B-A3B: 37.5 (#194)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| SciCode | 38.9% | 33.3% |
| WeirdML | 39.5% | 29.8% |
| LMArena Coding | 1454 | 1416 |
| SWE-bench Verified (bash only) | 70% | — |
| Aider Polyglot | 74.2% | — |
| LMArena WebDev | 1362 | — |
| SWE-bench Multilingual | 59% | — |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), Qwen3-30B-A3B: 29.8 (#82)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 56.7% | 41.4% |
| Terminal-Bench | 39.6% | — |
| APEX-Agents | 21.3% | — |
| TheAgentCompany | 42.9% | — |
| Vending-Bench 2 | 1,034 | — |
Reasoning Too close to call
DeepSeek-V3.2-Exp: 22.1 (#208), Qwen3-30B-A3B: 22.2 (#204)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 54.9% |
| CritPt | 2.9% | 0.3% |
| Chess Puzzles | 14% | 8% |
| LMArena Hard Prompts | 1434 | 1398 |
| DTBench | 87.7% | 69.3% |
| LMCA | 29.1% | 22.4% |
| Epoch Capabilities Index | 146.27 | 139.63 |
| ARC-AGI-2 | 4% | — |
| NYT Connections (extended) | 36.7% | — |
| ARC-AGI-1 | 57% | — |
| Thematic Generalization | 65% | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), Qwen3-30B-A3B: 37.4 (#157)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| MathArena Final-Answer Competitions | 57.7% | 47.8% |
| OTIS Mock AIME 2024-2025 | 87.8% | 70.3% |
| LMArena Math | 1435 | 1394 |
| ProofBench | 8% | — |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), Qwen3-30B-A3B: 41.8 (#105)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| GPQA Diamond | 83.4% | 70.1% |
| LMArena Expert | 1436 | 1396 |
| Confabulations | — | 12.3% |
| Vectara Hallucination Rate | 5.3% | — |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), Qwen3-30B-A3B: 49.5 (#132)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1409 | 1372 |
| LMArena Chinese | 1461 | 1433 |
| LMArena French | 1433 | 1418 |
| LMArena German | 1440 | 1380 |
| LMArena Japanese | 1374 | 1337 |
| LMArena Korean | 1371 | 1331 |
| LMArena Russian | 1424 | 1370 |
| LMArena Spanish | 1440 | 1404 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), Qwen3-30B-A3B: 72.0 (#142)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1413 | 1363 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), Qwen3-30B-A3B: 31.0 (#283)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| Fiction.LiveBench | 83.3% | 40.6% |
| LMArena Longer Query | 1428 | 1379 |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), Qwen3-30B-A3B: 55.6 (#143)
| Benchmark | DeepSeek-V3.2-Exp | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1425 | 1384 |
| LMArena Creative Writing | 1403 | 1317 |
| LMArena Multi-Turn | 1427 | 1378 |
| Short-Story Creative Writing | — | 75.3% |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than Qwen3-30B-A3B?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 38.9 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or Qwen3-30B-A3B?
Qwen3-30B-A3B is cheaper. It lists at $0.12 per million input tokens and $0.50 per million output tokens; DeepSeek-V3.2-Exp lists at $0.26 and $0.38.
Is DeepSeek-V3.2-Exp or Qwen3-30B-A3B better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 37.5 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3.2-Exp does, with 164K tokens against 41K.
How many benchmarks do DeepSeek-V3.2-Exp and Qwen3-30B-A3B share?
30 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Qwen3-30B-A3B has 32.