Model comparison
DeepSeek-R1 vs Qwen2.5 7B Instruct
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 29.0 on the Noometry Index. Qwen2.5 7B Instruct costs 3.0× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 9 shared benchmarks.
DeepSeek
Alibaba (Qwen)Summary
- They share 9 benchmarks with published results for both. DeepSeek-R1 scores higher in 7 categories and Qwen2.5 7B Instruct in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 12.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 2.5% for Qwen2.5 7B Instruct.
- Qwen2.5 7B Instruct is cheaper at $0.17 / $0.70 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- DeepSeek-R1 accepts more context: 164K tokens versus 131K.
- Qwen2.5 7B Instruct has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Qwen2.5 7B Instruct | |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.3 | 29.0 |
| Released | 2025-01-20 | 2024-09 |
| Weights | Proprietary | Open |
| Context window | 164K | 131K |
| Max output | 64K | 8K |
| Input $ / M tokens | $0.50 | $0.17 |
| Output $ / M tokens | $2.15 | $0.70 |
| Results tracked | 52 | 15 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Qwen2.5 7B Instruct: 36.5 (#208)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| BigCodeBench Instruct | — | 37.6% |
| LiveBench Coding | 66.7% | — |
| LMArena Coding | 1427 | — |
| BigCodeBench Complete | — | 46.1% |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), Qwen2.5 7B Instruct: 23.8 (#124)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| BALROG | 34.9% | 7.8% |
| DeepResearch Bench | 35.1% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Qwen2.5 7B Instruct: 14.8 (#322)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| Epoch Capabilities Index | 141.29 | 118.51 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 83.2% | — |
| LMArena Hard Prompts | 1416 | — |
| DTBench | — | 47.7% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 6.4% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Qwen2.5 7B Instruct: 12.6 (#306)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 2.5% |
| Omni-MATH | 42.4% | 29.4% |
| LiveBench Math | 80.7% | — |
| LMArena Math | 1400 | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Qwen2.5 7B Instruct: 17.0 (#286)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| GPQA Diamond | 76.3% | 35.5% |
| MMLU-Pro | 79.3% | 53.9% |
| GPQA (HELM) | 66.6% | 34.1% |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| LMArena Expert | 1394 | — |
| MMLU | — | 72.9% |
Multilingual Not comparable
DeepSeek-R1: 52.4 (#85), Qwen2.5 7B Instruct: —
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| LMArena Non-English | 1412 | — |
| LMArena Chinese | 1442 | — |
| LMArena French | 1417 | — |
| LMArena German | 1404 | — |
| LMArena Japanese | 1391 | — |
| LMArena Korean | 1360 | — |
| LMArena Russian | 1423 | — |
| LMArena Spanish | 1411 | — |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), Qwen2.5 7B Instruct: 63.2 (#231)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| IFEval | 78.4% | 74.1% |
| LiveBench Instruction Following | 80.5% | — |
| LMArena Instruction Following | 1382 | — |
Long Context Not comparable
DeepSeek-R1: 45.4 (#36), Qwen2.5 7B Instruct: —
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| Fiction.LiveBench | 75% | — |
| LMArena Longer Query | 1391 | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Qwen2.5 7B Instruct: 48.8 (#195)
| Benchmark | DeepSeek-R1 | Qwen2.5 7B Instruct |
|---|---|---|
| WildBench | 82.8% | 73.1% |
| LMArena Text | 1428 | — |
| LMArena Creative Writing | 1405 | — |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| LMArena Multi-Turn | 1405 | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Qwen2.5 7B Instruct?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 29.0 on the Noometry Index. Qwen2.5 7B Instruct costs 3.0× 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 Qwen2.5 7B Instruct?
Qwen2.5 7B Instruct is cheaper. It lists at $0.17 per million input tokens and $0.70 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or Qwen2.5 7B Instruct better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 36.5 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-R1 and Qwen2.5 7B Instruct share?
9 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen2.5 7B Instruct has 15.