Llama 3.1 8B Instruct

Meta llm

Meta's 8B-parameter multilingual instruction-tuned model from the Llama 3.1 family, released alongside the 70B and 405B variants.

overall ⌄

25.0

93 of 93 ranked

price · $/M tokens

self-hosted

open weights, no first-party API

to run it yourself ⌄

5 GB

minimum · 19 GB recommended

Benchmarks

What its maker published, and what anyone else measured. Every figure links the document it came from.

benchmark	published	measured
MMLU-Pro	48.3	—
GPQA	31.8	—

↑ finetune Llama 3.1 8B