Model comparison
DeepSeek-R1 vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index.
Last verified . 23 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. DeepSeek-R1 scores higher in 3 categories and Qwen3.8 27B in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.8 27B leads 41.0 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-1: 21.2% for DeepSeek-R1 and 87.5% for Qwen3.8 27B.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $0.99 / $1.49 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 164K.
- Qwen3.8 27B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Qwen3.8 27B | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.3 | 46.0 |
| Released | 2025-01-20 | 2026-08-14 |
| Weights | Proprietary | Open |
| Context window | 164K | 262K |
| Max output | 64K | 33K |
| Input $ / M tokens | $0.50 | $0.99 |
| Output $ / M tokens | $2.15 | $1.49 |
| Results tracked | 52 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
DeepSeek-R1: 46.3 (#68), Qwen3.8 27B: 50.5 (#44)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| SciCode | 35.7% | 46.6% |
| LMArena Coding | 1427 | 1482 |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1593 |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Qwen3.8 27B leads
DeepSeek-R1: 30.7 (#75), Qwen3.8 27B: 32.9 (#57)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning Qwen3.8 27B leads
DeepSeek-R1: 18.6 (#278), Qwen3.8 27B: 41.0 (#54)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| ARC-AGI-2 | 1.3% | 42.4% |
| ARC-AGI-1 | 21.2% | 87.5% |
| CritPt | 1.1% | 5.4% |
| LMArena Hard Prompts | 1416 | 1460 |
| Epoch Capabilities Index | 141.29 | 149.38 |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 54.5% |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 88% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Qwen3.8 27B: 37.1 (#161)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1400 | 1456 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| ProofBench | — | 16% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Qwen3.8 27B: 41.6 (#109)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1394 | 1482 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
DeepSeek-R1: 52.4 (#85), Qwen3.8 27B: 53.7 (#60)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1412 | 1430 |
| LMArena Chinese | 1442 | 1504 |
| LMArena French | 1417 | 1465 |
| LMArena German | 1404 | 1438 |
| LMArena Japanese | 1391 | 1384 |
| LMArena Korean | 1360 | 1393 |
| LMArena Russian | 1423 | 1415 |
| LMArena Spanish | 1411 | 1448 |
Instruction Following Qwen3.8 27B leads
DeepSeek-R1: 72.0 (#143), Qwen3.8 27B: 75.8 (#53)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1382 | 1439 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Qwen3.8 27B: 44.3 (#70)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1391 | 1450 |
| Fiction.LiveBench | 75% | — |
Writing & Preference Qwen3.8 27B leads
DeepSeek-R1: 61.4 (#88), Qwen3.8 27B: 65.8 (#43)
| Benchmark | DeepSeek-R1 | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1428 | 1441 |
| LMArena Creative Writing | 1405 | 1384 |
| EQ-Bench Creative Writing | 1500 | 1671 |
| LMArena Multi-Turn | 1405 | 1441 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or Qwen3.8 27B?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Qwen3.8 27B lists at $0.99 and $1.49.
Is DeepSeek-R1 or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-R1 and Qwen3.8 27B share?
23 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen3.8 27B has 31.