Model comparison
DeepSeek-R1 vs Qwen3 Max
Qwen3 Max is the stronger model overall, scoring 43.7 to 42.3 on the Noometry Index. DeepSeek-R1 costs 2.6× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. DeepSeek-R1 scores higher in 3 categories and Qwen3 Max in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 38.7.
- The biggest single-benchmark swing is Fiction.LiveBench: 75% for DeepSeek-R1 and 66.7% for Qwen3 Max.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $1.20 / $6 for Qwen3 Max.
- Qwen3 Max accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-R1 | Qwen3 Max | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.3 | 43.7 |
| Released | 2025-01-20 | 2025-09-23 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 262K |
| Max output | 64K | 66K |
| Input $ / M tokens | $0.50 | $1.20 |
| Output $ / M tokens | $2.15 | $6 |
| Results tracked | 52 | 33 |
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 Max: 43.0 (#93)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| LMArena Coding | 1427 | 1456 |
| ALE-Bench | 804.12 | 370.45 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Not comparable
DeepSeek-R1: 30.7 (#75), Qwen3 Max: —
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 71.56 |
Reasoning Qwen3 Max leads
DeepSeek-R1: 18.6 (#278), Qwen3 Max: 22.6 (#190)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| Kagi LLM Benchmark | 69.4% | 72.5% |
| LMArena Hard Prompts | 1416 | 1448 |
| Epoch Capabilities Index | 141.29 | 142.38 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| NYT Connections (extended) | — | 30.1% |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 4% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 5% |
| DTBench | — | 82.1% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 28.3% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Qwen3 Max: 38.7 (#131)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 73.3% |
| LMArena Math | 1400 | 1446 |
| MATH Level 5 | 96.6% | 97.1% |
| FrontierMath (Tiers 1-3) | — | 18.9% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
Knowledge Qwen3 Max leads
DeepSeek-R1: 44.5 (#87), Qwen3 Max: 48.1 (#78)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 76.3% | 72.6% |
| LMArena Expert | 1394 | 1455 |
| SimpleQA Verified | — | 48.7% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multilingual Qwen3 Max leads
DeepSeek-R1: 52.4 (#85), Qwen3 Max: 53.7 (#62)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| LMArena Non-English | 1412 | 1429 |
| LMArena Chinese | 1442 | 1478 |
| LMArena French | 1417 | 1449 |
| LMArena German | 1404 | 1463 |
| LMArena Japanese | 1391 | 1397 |
| LMArena Korean | 1360 | 1399 |
| LMArena Russian | 1423 | 1428 |
| LMArena Spanish | 1411 | 1462 |
Instruction Following Qwen3 Max leads
DeepSeek-R1: 72.0 (#143), Qwen3 Max: 74.8 (#87)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| LMArena Instruction Following | 1382 | 1419 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Qwen3 Max: 41.6 (#134)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| Fiction.LiveBench | 75% | 66.7% |
| LMArena Longer Query | 1391 | 1438 |
| CL-bench | — | 14.5% |
Writing & Preference Qwen3 Max leads
DeepSeek-R1: 61.4 (#88), Qwen3 Max: 62.4 (#76)
| Benchmark | DeepSeek-R1 | Qwen3 Max |
|---|---|---|
| LMArena Text | 1428 | 1439 |
| LMArena Creative Writing | 1405 | 1402 |
| LMArena Multi-Turn | 1405 | 1446 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Qwen3 Max?
Qwen3 Max is the stronger model overall, scoring 43.7 to 42.3 on the Noometry Index. DeepSeek-R1 costs 2.6× less per token, which makes it the better buy when Qwen3 Max's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or Qwen3 Max?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Qwen3 Max lists at $1.20 and $6.
Is DeepSeek-R1 or Qwen3 Max better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 43.0 in the Noometry coding category.
Which has the bigger context window?
Qwen3 Max does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-R1 and Qwen3 Max share?
24 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen3 Max has 33.