Llama 3 8B: Complete Benchmark Performance Guide

meta 3-8b 2024-04-18 Llama 3 Community License

In-depth analysis of Llama 3 8B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendormeta
Version3-8b
Release Date2024-04-18
Context Window8192 tokens
Input Modalitiestext
Output Modalitiestext
LicenseLlama 3 Community License
Documentationhttps://llama.meta.com/docs/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 67.8 % 2024-04-18 5-shot view
HUMANEVAL 65 pass@1 2024-04-18 — view
GSM8K 56.6 % 2024-04-18 0-shot CoT view
MATH 40 % 2024-04-18 0-shot CoT view
BBH 65 % 2024-04-18 3-shot CoT view
GPQA 27.9 % 2024-04-18 0-shot view
IFEVAL 68.2 % 2024-04-18 prompt_strict view
ARC 91.2 % 2024-04-18 challenge view
MUSR 46.1 % 2024-04-18 0-shot view
WINOGRANDE 75.6 % 2024-04-18 0-shot view

Pricing

Tier Price Currency
Input$0.18 / MtokUSD
Output$0.18 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://ai.meta.com/blog/ · as of 2024-04-18

Compliance

  • Data Residency: self-host
  • SOC2: ✗
  • HIPAA: ✗
  • GDPR: ✗
  • ISO 27001: ✗

Llama 3 8B

Model Overview

Meta Llama 3 8B 经济型开源模型, 8K 上下文, 8B 参数, 性能超越 Llama 2 13B, 适合本地部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Meta 3-8b 2024-04-18 8K text text Llama 3 Community License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 67.8 % 5-shot
HumanEval 65.0 pass@1 —
GSM8K (Grade School Math 8K) 56.6 % 0-shot CoT
MATH 40.0 % 0-shot CoT
BBH (BIG-Bench Hard) 65.0 % 3-shot CoT
GPQA 27.9 % 0-shot
IFEval 68.2 % prompt_strict
ARC 91.2 % challenge
MUSR 46.1 % 0-shot
WinoGrande 75.6 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • Proprietary, not self-hostable.
  • Context window 8K is limited.

Use Cases

  • General chat and Q&A

References