Model comparison
DeepSeek-R1 vs Qwen3-30B-A3B
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 38.9 on the Noometry Index. Qwen3-30B-A3B costs 4.2× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 27 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. DeepSeek-R1 scores higher in 7 categories and Qwen3-30B-A3B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in long context, where DeepSeek-R1 leads 45.4 to 31.0.
- The biggest single-benchmark swing is Fiction.LiveBench: 75% for DeepSeek-R1 and 40.6% for Qwen3-30B-A3B.
- Qwen3-30B-A3B is cheaper at $0.12 / $0.50 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- DeepSeek-R1 accepts more context: 164K tokens versus 41K.
- Qwen3-30B-A3B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Qwen3-30B-A3B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.3 | 38.9 |
| Released | 2025-01-20 | 2025-04-28 |
| Weights | Proprietary | Open |
| Context window | 164K | 41K |
| Max output | 64K | 16K |
| Input $ / M tokens | $0.50 | $0.12 |
| Output $ / M tokens | $2.15 | $0.50 |
| Results tracked | 52 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Qwen3-30B-A3B: 37.5 (#194)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| SciCode | 35.7% | 33.3% |
| WeirdML | 41.6% | 29.8% |
| LMArena Coding | 1427 | 1416 |
| Aider Polyglot | 71.4% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Too close to call
DeepSeek-R1: 30.7 (#75), Qwen3-30B-A3B: 29.8 (#82)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 41.4% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning Qwen3-30B-A3B leads
DeepSeek-R1: 18.6 (#278), Qwen3-30B-A3B: 22.2 (#204)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| Kagi LLM Benchmark | 69.4% | 54.9% |
| CritPt | 1.1% | 0.3% |
| LMArena Hard Prompts | 1416 | 1398 |
| Epoch Capabilities Index | 141.29 | 139.63 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| ARC-AGI-1 | 21.2% | — |
| Chess Puzzles | — | 8% |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 69.3% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 22.4% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Qwen3-30B-A3B: 37.4 (#157)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 70.3% |
| LMArena Math | 1400 | 1394 |
| MathArena Final-Answer Competitions | — | 47.8% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Qwen3-30B-A3B: 41.8 (#105)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| GPQA Diamond | 76.3% | 70.1% |
| Confabulations | 12.7% | 12.3% |
| LMArena Expert | 1394 | 1396 |
| MMLU-Pro | 79.3% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Qwen3-30B-A3B: 49.5 (#132)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1412 | 1372 |
| LMArena Chinese | 1442 | 1433 |
| LMArena French | 1417 | 1418 |
| LMArena German | 1404 | 1380 |
| LMArena Japanese | 1391 | 1337 |
| LMArena Korean | 1360 | 1331 |
| LMArena Russian | 1423 | 1370 |
| LMArena Spanish | 1411 | 1404 |
Instruction Following Too close to call
DeepSeek-R1: 72.0 (#143), Qwen3-30B-A3B: 72.0 (#142)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1382 | 1363 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Qwen3-30B-A3B: 31.0 (#283)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| Fiction.LiveBench | 75% | 40.6% |
| LMArena Longer Query | 1391 | 1379 |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Qwen3-30B-A3B: 55.6 (#143)
| Benchmark | DeepSeek-R1 | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1428 | 1384 |
| LMArena Creative Writing | 1405 | 1317 |
| Short-Story Creative Writing | 83% | 75.3% |
| LMArena Multi-Turn | 1405 | 1378 |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Qwen3-30B-A3B?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 38.9 on the Noometry Index. Qwen3-30B-A3B costs 4.2× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 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-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or Qwen3-30B-A3B better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 37.5 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 41K.
How many benchmarks do DeepSeek-R1 and Qwen3-30B-A3B share?
27 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen3-30B-A3B has 32.