MA
Meta AI
Language model · United States ·Jul 2024

Llama 3.1-8B

Language Open (Restricted)
8B
Parameters
67
Benchmarks

About Llama 3.1-8B

Llama 3.1-8B is an AI model developed by Meta AI, in the language category, released in 2024, made available as an open-weights model with 8B params.

On this page you'll find Llama 3.1-8B's full specifications, including 67 benchmark results. Review provider pricing and benchmark scores below, or compare Llama 3.1-8B head-to-head with other language models.

Links

Run Llama 3.1-8B locally — what to buy

Share this build ↗

Sized to what Llama 3.1-8B actually needs (~8k context at Q4), not to the biggest GPU: Good = cheapest that runs it, Better = best value, Best = most headroom. Speed is a hardware estimate; anything past ~120 tok/s is shown as instant.

Software support

✓ measured · · compatible · — not supported. Informational only — speed is hardware-based.

Community

Would you run Llama 3.1-8B again?
No votes yet — be the first.
Did it run? — community reports by hardware

No reports yet. Own this model on some hardware? Be the first to confirm it runs.

Discussion

No comments yet. Start the discussion.

Llama 3.1-8B benchmark scores
Benchmark Category Score Variant
HumanEval code 86.3% instruct-pass@10-cited-tulu3
HumanEval code 85.32% instruct-cited-granite-3-2
HumanEval code 66.5% instruct-cited-phi35moe
HumanEval+ code 82.9% instruct-pass@10-cited-tulu3
HumanEval+ code 80.15% instruct-cited-granite-3-2
MBPP code 69.4% instruct-cited-phi35moe
AlpacaEval 2 general 27.22% instruct-cited-granite-3-2
AlpacaEval 2 general 25.7% instruct-cited-exaone35
AlpacaEval 2 general 24.2% instruct-lc-win-cited-tulu3
ArenaHard general 36.43% instruct-cited-granite-3-2
ArenaHard general 27.7% instruct-cited-exaone35
ArenaHard general 25.7% instruct-cited-phi35moe
IFEval general 80.6% instruct-cited-tulu3
IFEval general 79.1% instruct-cited-granite-3-2
IFEval general 78.8% instruct-cited-falcon3
IFEval general 12.7% base-cited-falcon-mamba
LiveBench general 28.3 instruct-cited-exaone35
MMLU general 71.2% instruct-cited-tulu3
MMLU general 69.15% instruct-cited-granite-3-2
MMLU general 68.2% instruct-5-shot-cited-falcon3
MMLU general 68.1% instruct-cited-phi35moe
MMLU general 66.9% base-cited-olmo2
MMLU general 66.4% base-5-shot-cited-marin
MMLU-Pro general 44% instruct-cited-phi35moe
MMLU-Pro general 36.4% instruct-5-shot-cited-falcon3
MMLU-Pro general 34.7% base-cited-olmo2
MMLU-Pro general 33.3% base-cited-marin
MMLU-Pro general 24.95% base-cited-falcon-mamba
MMMLU general 56.2% instruct-cited-phi35moe
AGIEval knowledge 51.3 base-cited-olmo2
ARC-Challenge knowledge 79.5 base-cited-olmo2
ARC-Challenge knowledge 58.9 base-cited-marin
ARC-Challenge knowledge 58.6 instruct-25-shot-cited-falcon3
ARC-Easy knowledge 85.8 base-cited-marin
TriviaQA knowledge 80.3 base-cited-olmo2
GSM8K math 83.4% instruct-cited-tulu3
GSM8K math 83.24% instruct-cited-granite-3-2
GSM8K math 82.6% instruct-5-shot-cited-falcon3
GSM8K math 56.5% base-cited-olmo2
GSM8K math 47.92% base-cited-falcon-mamba
MATH math 47.6% instruct-cited-phi35moe
MATH math 42.5% instruct-cited-tulu3
MGSM math 56.7 instruct-cited-phi35moe
BIG-Bench Hard reasoning 72.66% instruct-cited-granite-3-2
BIG-Bench Hard reasoning 63.4% instruct-cited-phi35moe
BIG-Bench Hard reasoning 62.8% instruct-cited-tulu3
BIG-Bench Hard reasoning 48.6% instruct-3-shot-cited-falcon3
BIG-Bench Hard reasoning 46.4% base-cited-marin
BIG-Bench Hard reasoning 25.29% base-cited-falcon-mamba
DROP reasoning 61.5% instruct-cited-tulu3
DROP reasoning 61.48% instruct-cited-granite-3-2
DROP reasoning 56.4% base-cited-olmo2
GPQA Diamond reasoning 33.5% instruct-0-shot-cited-falcon3
GPQA Diamond reasoning 26.3% instruct-cited-phi35moe
GPQA Diamond reasoning 6.15% base-cited-falcon-mamba
HellaSwag reasoning 81.9 base-10-shot-cited-marin
HellaSwag reasoning 81.6 base-cited-olmo2
PIQA reasoning 81.2 instruct-cited-phi35moe
PIQA reasoning 78.9 instruct-0-shot-cited-falcon3
PIQA reasoning 45.8 base-cited-marin
Winogrande reasoning 82.9 base-cited-marin
Winogrande reasoning 76.6 base-cited-olmo2
Winogrande reasoning 64.7 instruct-cited-phi35moe
TruthfulQA safety 69.2% instruct-cited-phi35moe
TruthfulQA safety 55.1% instruct-cited-tulu3
TruthfulQA safety 52.79% instruct-cited-granite-3-2
TruthfulQA safety 44.29% base-cited-falcon-mamba

More models like Llama 3.1-8B

Frequently asked questions

What is Llama 3.1-8B?

Llama 3.1-8B is an AI model developed by Meta AI, in the language category, released in 2024. It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.

Who created Llama 3.1-8B?

Llama 3.1-8B was developed by Meta AI and released in 2024.

Is Llama 3.1-8B open source or proprietary?

Llama 3.1-8B ships as an open-weights model: you can download and self-host the weights, though the licence may place some restrictions on use.

How much does Llama 3.1-8B cost?

Pricing for Llama 3.1-8B depends on the provider. See the providers table on this page for the latest API rates.

How does Llama 3.1-8B perform on benchmarks?

Llama 3.1-8B is benchmarked across 67 evaluations on GenAIList, including HumanEval (86.3%). See the full benchmark table below and compare it with other models.

No videos yet. Know a good one? Add it below.

Know a great video about Llama 3.1-8B?