Model comparison
DeepSeek-V3.1 vs Qwen3-30B-A3B
DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 38.5 on the Noometry Index. Qwen3-30B-A3B costs 2.0× less per token, which makes it the better buy when DeepSeek-V3.1's lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 8 categories and Qwen3-30B-A3B in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.1 leads 27.3 to 21.6.
- The biggest single-benchmark swing is DTBench: 82.7% for DeepSeek-V3.1 and 69.3% for Qwen3-30B-A3B.
- Qwen3-30B-A3B is cheaper at $0.12 / $0.50 per million input/output tokens, against $0.25 / $0.95 for DeepSeek-V3.1.
- DeepSeek-V3.1 accepts more context: 164K tokens versus 41K.
Side by side
| DeepSeek-V3.1 | Qwen3-30B-A3B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.2 | 38.5 |
| Released | 2025-08-21 | 2025-04-28 |
| Weights | Open | Open |
| Context window | 164K | 41K |
| Max output | 8K | 16K |
| Input $ / M tokens | $0.25 | $0.12 |
| Output $ / M tokens | $0.95 | $0.50 |
| Results tracked | 27 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1 leads
DeepSeek-V3.1: 39.8 (#140), Qwen3-30B-A3B: 37.0 (#188)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| WeirdML | 38.4% | 29.8% |
| LMArena Coding | 1417 | 1416 |
| SciCode | — | 33.3% |
Agentic & Tool Use Not comparable
DeepSeek-V3.1: —, Qwen3-30B-A3B: 29.5 (#75)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 41.4% |
Reasoning DeepSeek-V3.1 leads
DeepSeek-V3.1: 27.3 (#109), Qwen3-30B-A3B: 21.6 (#193)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 54.9% |
| LMArena Hard Prompts | 1417 | 1398 |
| DTBench | 82.7% | 69.3% |
| LMCA | 24.3% | 22.4% |
| Epoch Capabilities Index | 139.92 | 139.63 |
| SimpleBench | 40% | — |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 8% |
| ForecastBench | 58 | — |
Math Too close to call
DeepSeek-V3.1: 36.1 (#125), Qwen3-30B-A3B: 36.0 (#133)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Math | 1420 | 1394 |
| MathArena Final-Answer Competitions | — | 47.8% |
| OTIS Mock AIME 2024-2025 | — | 70.3% |
Knowledge DeepSeek-V3.1 leads
DeepSeek-V3.1: 43.0 (#86), Qwen3-30B-A3B: 41.7 (#96)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Expert | 1403 | 1395 |
| GPQA Diamond | — | 70.1% |
| Confabulations | — | 12.3% |
| Vectara Hallucination Rate | 5.5% | — |
Multilingual DeepSeek-V3.1 leads
DeepSeek-V3.1: 51.8 (#100), Qwen3-30B-A3B: 49.8 (#126)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1400 | 1372 |
| LMArena Chinese | 1470 | 1434 |
| LMArena French | 1447 | 1419 |
| LMArena German | 1412 | 1379 |
| LMArena Japanese | 1374 | 1338 |
| LMArena Korean | 1337 | 1333 |
| LMArena Russian | 1405 | 1369 |
| LMArena Spanish | 1429 | 1403 |
Instruction Following DeepSeek-V3.1 leads
DeepSeek-V3.1: 74.1 (#105), Qwen3-30B-A3B: 72.2 (#137)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1400 | 1364 |
Long Context DeepSeek-V3.1 leads
DeepSeek-V3.1: 36.3 (#227), Qwen3-30B-A3B: 31.0 (#278)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| Fiction.LiveBench | 52.8% | 40.6% |
| LMArena Longer Query | 1422 | 1379 |
Writing & Preference DeepSeek-V3.1 leads
DeepSeek-V3.1: 60.5 (#93), Qwen3-30B-A3B: 55.9 (#138)
| Benchmark | DeepSeek-V3.1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1420 | 1384 |
| LMArena Creative Writing | 1401 | 1318 |
| LMArena Multi-Turn | 1408 | 1378 |
| Short-Story Creative Writing | — | 75.3% |
| EQ-Bench Creative Writing | 1436 | — |
Frequently asked questions
Is DeepSeek-V3.1 better than Qwen3-30B-A3B?
DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 38.5 on the Noometry Index. Qwen3-30B-A3B costs 2.0× less per token, which makes it the better buy when DeepSeek-V3.1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3.1 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-V3.1 lists at $0.25 and $0.95.
Is DeepSeek-V3.1 or Qwen3-30B-A3B better for coding?
DeepSeek-V3.1 scores higher on coding benchmarks: 39.8 versus 37.0 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-V3.1 does, with 164K tokens against 41K.
How many benchmarks do DeepSeek-V3.1 and Qwen3-30B-A3B share?
23 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Qwen3-30B-A3B has 32.