Model comparison
DeepSeek-R1-Distill-Qwen-14B vs DeepSeek-R1-Distill-Qwen-32B
DeepSeek-R1-Distill-Qwen-32B is the stronger model overall, scoring 35.5 to 32.7 on the Noometry Index.
Last verified . 6 shared benchmarks.
DeepSeekSummary
- They share 6 benchmarks with published results for both. DeepSeek-R1-Distill-Qwen-14B scores higher in 3 categories and DeepSeek-R1-Distill-Qwen-32B in 1 category; one gap is clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1-Distill-Qwen-32B leads 35.7 to 24.1.
- The biggest single-benchmark swing is GPQA Diamond: 44.7% for DeepSeek-R1-Distill-Qwen-14B and 64.1% for DeepSeek-R1-Distill-Qwen-32B.
Side by side
| DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B | |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Noometry Index | 32.7 | 35.5 |
| Released | 2025-01-20 | 2025-01-20 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 7 | 14 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-R1-Distill-Qwen-14B: 36.9 (#200), DeepSeek-R1-Distill-Qwen-32B: 36.1 (#212)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| BigCodeBench Instruct | 38.1% | 43.9% |
| BigCodeBench Complete | 48.4% | 54.9% |
| LiveBench Coding | — | 33.7% |
Agentic & Tool Use Not comparable
DeepSeek-R1-Distill-Qwen-14B: —, DeepSeek-R1-Distill-Qwen-32B: 28.1 (#94)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| BALROG | — | 19.5% |
Reasoning Too close to call
DeepSeek-R1-Distill-Qwen-14B: 19.2 (#263), DeepSeek-R1-Distill-Qwen-32B: 18.2 (#284)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| Chess Puzzles | 1% | 1% |
| Epoch Capabilities Index | 135.43 | 137.44 |
| LiveBench Reasoning | — | 52.3% |
| LiveBench Data Analysis | — | 45.4% |
| LiveBench | — | 45.5% |
Math Too close to call
DeepSeek-R1-Distill-Qwen-14B: 35.5 (#184), DeepSeek-R1-Distill-Qwen-32B: 34.5 (#194)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 50.6% | 55.6% |
| LiveBench Math | — | 59.4% |
| MATH Level 5 | 87.1% | — |
Knowledge DeepSeek-R1-Distill-Qwen-32B leads
DeepSeek-R1-Distill-Qwen-14B: 24.1 (#270), DeepSeek-R1-Distill-Qwen-32B: 35.7 (#182)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| GPQA Diamond | 44.7% | 64.1% |
Instruction Following Not comparable
DeepSeek-R1-Distill-Qwen-14B: —, DeepSeek-R1-Distill-Qwen-32B: 61.6 (#243)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| LiveBench Instruction Following | — | 55.7% |
Writing & Preference Not comparable
DeepSeek-R1-Distill-Qwen-14B: —, DeepSeek-R1-Distill-Qwen-32B: 49.6 (#188)
| Benchmark | DeepSeek-R1-Distill-Qwen-14B | DeepSeek-R1-Distill-Qwen-32B |
|---|---|---|
| LiveBench Language | — | 26.8% |
Frequently asked questions
Is DeepSeek-R1-Distill-Qwen-14B better than DeepSeek-R1-Distill-Qwen-32B?
DeepSeek-R1-Distill-Qwen-32B is the stronger model overall, scoring 35.5 to 32.7 on the Noometry Index.
Is DeepSeek-R1-Distill-Qwen-14B or DeepSeek-R1-Distill-Qwen-32B better for coding?
They score almost the same on coding (36.9 vs 36.1); test both on your own repository before choosing.
How many benchmarks do DeepSeek-R1-Distill-Qwen-14B and DeepSeek-R1-Distill-Qwen-32B share?
6 benchmarks have published results for both models. DeepSeek-R1-Distill-Qwen-14B has 7 scored results on Noometry and DeepSeek-R1-Distill-Qwen-32B has 14.