Compare · ModelsLive · 2 picked · head to head
Llama 3-70B vs Llama 3 8B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Llama 3-70B wins on 14/15 benchmarks
Llama 3-70B wins 14 of 15 shared benchmarks. Leads in knowledge · general · language.
Category leads
knowledge·Llama 3-70Bgeneral·Llama 3-70Blanguage·Llama 3-70Bmath·Llama 3-70Breasoning·Llama 3 8B Instruct
Hype vs Reality
Attention vs performance
Llama 3-70B
#243 by perf·#18 by attention
Llama 3 8B Instruct
#248 by perf·#18 by attention
Best value
Llama 3 8B Instruct
Llama 3-70B
n/a
no price
Llama 3 8B Instruct
220.0 pts/$
$0.14/M
Vendor risk
Who is behind the model
Meta AI
$1.87T·Tier 1
Meta AI
$1.87T·Tier 1
Head to head
15 benchmarks · 2 models
Llama 3-70BLlama 3 8B Instruct
ANLI
ANLI (Adversarial NLI) · adversarially constructed natural language inference dataset where each round targets weaknesses found in previous model generations.
Llama 3-70B
36.0
Llama 3 8B Instruct
36.0
ARC AI2
AI2 Reasoning Challenge · tests grade-school level science knowledge with multiple-choice questions requiring reasoning beyond simple retrieval.
Llama 3-70B
77.1
Llama 3 8B Instruct
77.1
GPQA diamond
Llama 3-70B leads by +19.3
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Llama 3-70B
20.8
Llama 3 8B Instruct
1.4
BBH (HuggingFace)
Llama 3-70B leads by +8.3
Llama 3-70B
26.7
Llama 3 8B Instruct
18.4
GPQA
Llama 3-70B leads by +5.9
Llama 3-70B
8.1
Llama 3 8B Instruct
2.1
IFEval
Llama 3-70B leads by +9.2
Llama 3-70B
33.2
Llama 3 8B Instruct
24.0
MATH Level 5
Llama 3-70B leads by +1.8
Llama 3-70B
5.7
Llama 3 8B Instruct
3.9
MMLU-PRO
Llama 3-70B leads by +6.3
Llama 3-70B
24.0
Llama 3 8B Instruct
17.8
MUSR
Llama 3 8B Instruct leads by +12.8
Llama 3-70B
7.1
Llama 3 8B Instruct
19.9
MATH level 5
Llama 3-70B leads by +16.4
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
Llama 3-70B
22.6
Llama 3 8B Instruct
6.1
MMLU
Llama 3-70B leads by +14.0
Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge.
Llama 3-70B
72.4
Llama 3 8B Instruct
58.4
OpenBookQA
OpenBookQA · science questions that require combining a given core fact with broad common knowledge, mimicking an open-book exam setting.
Llama 3-70B
76.8
Llama 3 8B Instruct
76.8
OTIS Mock AIME 2024-2025
Llama 3-70B leads by +3.5
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Llama 3-70B
4.2
Llama 3 8B Instruct
0.7
TriviaQA
TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents.
Llama 3-70B
67.7
Llama 3 8B Instruct
67.7
Winogrande
Llama 3-70B leads by +15.6
WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs.
Llama 3-70B
67.0
Llama 3 8B Instruct
51.4
Full benchmark table
| Benchmark | Llama 3-70B | Llama 3 8B Instruct |
|---|---|---|
ANLI ANLI (Adversarial NLI) · adversarially constructed natural language inference dataset where each round targets weaknesses found in previous model generations. | 36.0 | 36.0 |
ARC AI2 AI2 Reasoning Challenge · tests grade-school level science knowledge with multiple-choice questions requiring reasoning beyond simple retrieval. | 77.1 | 77.1 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 20.8 | 1.4 |
BBH (HuggingFace) | 26.7 | 18.4 |
GPQA | 8.1 | 2.1 |
IFEval | 33.2 | 24.0 |
MATH Level 5 | 5.7 | 3.9 |
MMLU-PRO | 24.0 | 17.8 |
MUSR | 7.1 | 19.9 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 22.6 | 6.1 |
MMLU Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge. | 72.4 | 58.4 |
OpenBookQA OpenBookQA · science questions that require combining a given core fact with broad common knowledge, mimicking an open-book exam setting. | 76.8 | 76.8 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 4.2 | 0.7 |
TriviaQA TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents. | 67.7 | 67.7 |
Winogrande WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs. | 67.0 | 51.4 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| $0.14 | $0.14 | 8K tokens (~4 books) | $1.40 |