Compare · ModelsLive · 3 picked · head to head
DeepSeek R1 Distill Qwen 14B vs DeepSeek R1 Distill Qwen 32B vs Qwen2 VL 7B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek R1 Distill Qwen 32B wins on 7/14 benchmarks
DeepSeek R1 Distill Qwen 32B wins 7 of 14 shared benchmarks. Leads in language.
Category leads
general·DeepSeek R1 Distill Qwen 14Bknowledge·DeepSeek R1 Distill Qwen 14Blanguage·DeepSeek R1 Distill Qwen 32Bmath·DeepSeek R1 Distill Qwen 14Breasoning·DeepSeek R1 Distill Qwen 14B
Hype vs Reality
Attention vs performance
DeepSeek R1 Distill Qwen 14B
#107 by perf·no signal
DeepSeek R1 Distill Qwen 32B
#163 by perf·no signal
Qwen2 VL 7B Instruct
#145 by perf·#2 by attention
Best value
Pricing unknown
DeepSeek R1 Distill Qwen 14B
n/a
no price
DeepSeek R1 Distill Qwen 32B
n/a
no price
Qwen2 VL 7B Instruct
n/a
no price
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
DeepSeek
$3.4B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
14 benchmarks · 3 models
DeepSeek R1 Distill Qwen 14BDeepSeek R1 Distill Qwen 32BQwen2 VL 7B Instruct
BBH (HuggingFace)
DeepSeek R1 Distill Qwen 14B leads by +4.8
DeepSeek R1 Distill Qwen 14B
40.7
DeepSeek R1 Distill Qwen 32B
17.1
Qwen2 VL 7B Instruct
35.9
GPQA
DeepSeek R1 Distill Qwen 14B leads by +9.1
DeepSeek R1 Distill Qwen 14B
18.3
DeepSeek R1 Distill Qwen 32B
4.6
Qwen2 VL 7B Instruct
9.3
IFEval
Qwen2 VL 7B Instruct leads by +2.2
DeepSeek R1 Distill Qwen 14B
43.8
DeepSeek R1 Distill Qwen 32B
41.9
Qwen2 VL 7B Instruct
46.0
MATH Level 5
DeepSeek R1 Distill Qwen 14B leads by +37.2
DeepSeek R1 Distill Qwen 14B
57.0
DeepSeek R1 Distill Qwen 32B
17.1
Qwen2 VL 7B Instruct
19.9
MMLU-PRO
DeepSeek R1 Distill Qwen 32B leads by +0.2
DeepSeek R1 Distill Qwen 14B
40.7
DeepSeek R1 Distill Qwen 32B
41.0
Qwen2 VL 7B Instruct
34.4
MUSR
DeepSeek R1 Distill Qwen 14B leads by +12.6
DeepSeek R1 Distill Qwen 14B
28.7
DeepSeek R1 Distill Qwen 32B
16.1
Qwen2 VL 7B Instruct
13.6
JCommonsenseQA
DeepSeek R1 Distill Qwen 32B leads by +1.6
DeepSeek R1 Distill Qwen 14B
93.7
DeepSeek R1 Distill Qwen 32B
95.3
Qwen2 VL 7B Instruct
87.8
JMMLU
DeepSeek R1 Distill Qwen 32B leads by +8.8
DeepSeek R1 Distill Qwen 14B
63.4
DeepSeek R1 Distill Qwen 32B
72.2
Qwen2 VL 7B Instruct
56.3
JNLI
DeepSeek R1 Distill Qwen 32B leads by +3.2
DeepSeek R1 Distill Qwen 14B
82.4
DeepSeek R1 Distill Qwen 32B
85.6
Qwen2 VL 7B Instruct
74.4
JSQuAD
Qwen2 VL 7B Instruct leads by +0.1
DeepSeek R1 Distill Qwen 14B
89.8
DeepSeek R1 Distill Qwen 32B
89.2
Qwen2 VL 7B Instruct
89.9
LLM-JP · Overall
DeepSeek R1 Distill Qwen 32B leads by +2.5
DeepSeek R1 Distill Qwen 14B
56.8
DeepSeek R1 Distill Qwen 32B
59.3
Qwen2 VL 7B Instruct
53.0
Chess Puzzles
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
DeepSeek R1 Distill Qwen 14B
0.0
DeepSeek R1 Distill Qwen 32B
0.0
GPQA diamond
DeepSeek R1 Distill Qwen 32B leads by +25.9
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
DeepSeek R1 Distill Qwen 14B
26.3
DeepSeek R1 Distill Qwen 32B
52.2
OTIS Mock AIME 2024-2025
DeepSeek R1 Distill Qwen 32B leads by +5.0
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek R1 Distill Qwen 14B
50.5
DeepSeek R1 Distill Qwen 32B
55.5
Full benchmark table
| Benchmark | DeepSeek R1 Distill Qwen 14B | DeepSeek R1 Distill Qwen 32B | Qwen2 VL 7B Instruct |
|---|---|---|---|
BBH (HuggingFace) | 40.7 | 17.1 | 35.9 |
GPQA | 18.3 | 4.6 | 9.3 |
IFEval | 43.8 | 41.9 | 46.0 |
MATH Level 5 | 57.0 | 17.1 | 19.9 |
MMLU-PRO | 40.7 | 41.0 | 34.4 |
MUSR | 28.7 | 16.1 | 13.6 |
JCommonsenseQA | 93.7 | 95.3 | 87.8 |
JMMLU | 63.4 | 72.2 | 56.3 |
JNLI | 82.4 | 85.6 | 74.4 |
JSQuAD | 89.8 | 89.2 | 89.9 |
LLM-JP · Overall | 56.8 | 59.3 | 53.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 0.0 | 0.0 | — |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 26.3 | 52.2 | — |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 50.5 | 55.5 | — |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| — | — | — | — | |
| — | — | — | — |