Compare · ModelsLive · 3 picked · head to head
DeepSeek R1 Distill Qwen 14B vs Hermes 3 70B Instruct vs Llama 3 8B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek R1 Distill Qwen 14B wins on 6/9 benchmarks
DeepSeek R1 Distill Qwen 14B wins 6 of 9 shared benchmarks. Leads in knowledge · math · reasoning.
Category leads
general·Hermes 3 70B Instructknowledge·DeepSeek R1 Distill Qwen 14Blanguage·Hermes 3 70B Instructmath·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
Hermes 3 70B Instruct
#200 by perf·no signal
Llama 3 8B Instruct
#248 by perf·#18 by attention
Best value
Llama 3 8B Instruct
4.0x better value than Hermes 3 70B Instruct
DeepSeek R1 Distill Qwen 14B
n/a
no price
Hermes 3 70B Instruct
55.0 pts/$
$0.70/M
Llama 3 8B Instruct
220.0 pts/$
$0.14/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
nousresearch
private · undisclosed
Meta AI
$1.87T·Tier 1
Head to head
9 benchmarks · 3 models
DeepSeek R1 Distill Qwen 14BHermes 3 70B InstructLlama 3 8B Instruct
BBH (HuggingFace)
Hermes 3 70B Instruct leads by +13.1
DeepSeek R1 Distill Qwen 14B
40.7
Hermes 3 70B Instruct
53.8
Llama 3 8B Instruct
18.4
GPQA
DeepSeek R1 Distill Qwen 14B leads by +3.5
DeepSeek R1 Distill Qwen 14B
18.3
Hermes 3 70B Instruct
14.9
Llama 3 8B Instruct
2.1
IFEval
Hermes 3 70B Instruct leads by +32.8
DeepSeek R1 Distill Qwen 14B
43.8
Hermes 3 70B Instruct
76.6
Llama 3 8B Instruct
24.0
MATH Level 5
DeepSeek R1 Distill Qwen 14B leads by +36.0
DeepSeek R1 Distill Qwen 14B
57.0
Hermes 3 70B Instruct
21.0
Llama 3 8B Instruct
3.9
MMLU-PRO
Hermes 3 70B Instruct leads by +0.7
DeepSeek R1 Distill Qwen 14B
40.7
Hermes 3 70B Instruct
41.4
Llama 3 8B Instruct
17.8
MUSR
DeepSeek R1 Distill Qwen 14B leads by +5.3
DeepSeek R1 Distill Qwen 14B
28.7
Hermes 3 70B Instruct
23.4
Llama 3 8B Instruct
19.9
GPQA diamond
DeepSeek R1 Distill Qwen 14B leads by +24.8
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
Llama 3 8B Instruct
1.4
MATH level 5
DeepSeek R1 Distill Qwen 14B leads by +81.0
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
DeepSeek R1 Distill Qwen 14B
87.1
Llama 3 8B Instruct
6.1
OTIS Mock AIME 2024-2025
DeepSeek R1 Distill Qwen 14B leads by +49.8
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek R1 Distill Qwen 14B
50.5
Llama 3 8B Instruct
0.7
Full benchmark table
| Benchmark | DeepSeek R1 Distill Qwen 14B | Hermes 3 70B Instruct | Llama 3 8B Instruct |
|---|---|---|---|
BBH (HuggingFace) | 40.7 | 53.8 | 18.4 |
GPQA | 18.3 | 14.9 | 2.1 |
IFEval | 43.8 | 76.6 | 24.0 |
MATH Level 5 | 57.0 | 21.0 | 3.9 |
MMLU-PRO | 40.7 | 41.4 | 17.8 |
MUSR | 28.7 | 23.4 | 19.9 |
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 | — | 1.4 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 87.1 | — | 6.1 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 50.5 | — | 0.7 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| $0.70 | $0.70 | 131K tokens (~66 books) | $7.00 | |
| $0.14 | $0.14 | 8K tokens (~4 books) | $1.40 |