Compare · ModelsLive · 2 picked · head to head
Meta Llama 3 8B Instruct vs DeepSeek R1 Distill Qwen 32B
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek R1 Distill Qwen 32B wins on 8/11 benchmarks
DeepSeek R1 Distill Qwen 32B wins 8 of 11 shared benchmarks. Leads in knowledge · language · math.
Category leads
general·Meta Llama 3 8B Instructknowledge·DeepSeek R1 Distill Qwen 32Blanguage·DeepSeek R1 Distill Qwen 32Bmath·DeepSeek R1 Distill Qwen 32Breasoning·DeepSeek R1 Distill Qwen 32B
Hype vs Reality
Attention vs performance
Meta Llama 3 8B Instruct
#142 by perf·no signal
DeepSeek R1 Distill Qwen 32B
#125 by perf·no signal
Best value
Pricing unknown
Meta Llama 3 8B Instruct
—
no price
DeepSeek R1 Distill Qwen 32B
—
no price
Vendor risk
Mixed exposure
One or more vendors flagged
Meta AI
$1.50T·Tier 1
DeepSeek
$3.4B·Tier 1
Head to head
11 benchmarks · 2 models
Meta Llama 3 8B InstructDeepSeek R1 Distill Qwen 32B
BBH (HuggingFace)
Meta Llama 3 8B Instruct leads by +11.1
Meta Llama 3 8B Instruct
28.2
DeepSeek R1 Distill Qwen 32B
17.1
GPQA
DeepSeek R1 Distill Qwen 32B leads by +3.4
Meta Llama 3 8B Instruct
1.2
DeepSeek R1 Distill Qwen 32B
4.6
IFEval
Meta Llama 3 8B Instruct leads by +32.2
Meta Llama 3 8B Instruct
74.1
DeepSeek R1 Distill Qwen 32B
41.9
MATH Level 5
DeepSeek R1 Distill Qwen 32B leads by +8.4
Meta Llama 3 8B Instruct
8.7
DeepSeek R1 Distill Qwen 32B
17.1
MMLU-PRO
DeepSeek R1 Distill Qwen 32B leads by +11.4
Meta Llama 3 8B Instruct
29.6
DeepSeek R1 Distill Qwen 32B
41.0
MUSR
DeepSeek R1 Distill Qwen 32B leads by +14.5
Meta Llama 3 8B Instruct
1.6
DeepSeek R1 Distill Qwen 32B
16.1
JCommonsenseQA
DeepSeek R1 Distill Qwen 32B leads by +7.6
Meta Llama 3 8B Instruct
87.7
DeepSeek R1 Distill Qwen 32B
95.3
JMMLU
DeepSeek R1 Distill Qwen 32B leads by +25.5
Meta Llama 3 8B Instruct
46.7
DeepSeek R1 Distill Qwen 32B
72.2
JNLI
DeepSeek R1 Distill Qwen 32B leads by +24.5
Meta Llama 3 8B Instruct
61.1
DeepSeek R1 Distill Qwen 32B
85.6
JSQuAD
Meta Llama 3 8B Instruct leads by +0.3
Meta Llama 3 8B Instruct
89.5
DeepSeek R1 Distill Qwen 32B
89.2
LLM-JP · Overall
DeepSeek R1 Distill Qwen 32B leads by +9.7
Meta Llama 3 8B Instruct
49.6
DeepSeek R1 Distill Qwen 32B
59.3
Full benchmark table
| Benchmark | Meta Llama 3 8B Instruct | DeepSeek R1 Distill Qwen 32B |
|---|---|---|
BBH (HuggingFace) | 28.2 | 17.1 |
GPQA | 1.2 | 4.6 |
IFEval | 74.1 | 41.9 |
MATH Level 5 | 8.7 | 17.1 |
MMLU-PRO | 29.6 | 41.0 |
MUSR | 1.6 | 16.1 |
JCommonsenseQA | 87.7 | 95.3 |
JMMLU | 46.7 | 72.2 |
JNLI | 61.1 | 85.6 |
JSQuAD | 89.5 | 89.2 |
LLM-JP · Overall | 49.6 | 59.3 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| — | — | — | — |