Meta: Llama 3.1 8B Instruct

Meta: Llama 3.1 8B Instruct

meta-llama · Released Jul 23, 2024
40
Our Score

Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors. This 8B instruct-tuned version is fast and efficient. It has demonstrated strong performance compared to leading closed-source models in human evaluations. Usage of this model is subject to Meta's Acceptable Use Policy.

$0.02 / 1M Input Price
$0.05 / 1M Output Price
16,384 tokens Context Window
16,384 tokens Max Output
8B Parameters

Capabilities

Tool Use Function Calling

Architecture

ModalityText → Text
TokenizerLlama3
Instruct Typellama3
Parameters8B

Performance Indices

Source: Artificial Analysis

11.8 Intelligence Index
4.9 Coding Index
8.6 Agentic Index
4.3 Math Index

Benchmark Scores

Evaluations

GPQA Diamond 25.9%
Graduate-level scientific reasoning
HLE 5.1%
Humanity's Last Exam
MMLU Pro 47.6%
Multi-task language understanding
LiveCodeBench 11.6%
Live coding evaluation
SciCode 13.2%
Scientific computing
MATH 500 51.9%
Mathematical problem-solving
AIME 7.7%
Competition mathematics
AIME 2025 4.3%
Competition mathematics (2025)
IFBench 28.6%
Instruction following
LCR 15.7%
Long-context reasoning
TerminalBench Hard 0.8%
Agentic terminal tasks
τ²-Bench 16.4%
Conversational agent benchmark

Benchmark data from Artificial Analysis and Hugging Face

Model Information

OpenRouter ID meta-llama/llama-3.1-8b-instruct
Providermeta-llama
Model FamilyLlama 3
Release Date July 23, 2024
Context Length16,384 tokens
Max Completion16,384 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.02 $0.000020
Output $0.05 $0.000050

Live Performance

Live endpoint metrics — refreshed every 30 minutes.

99.7%
Avg Uptime
64ms
Best Latency (TTFT)
282 tok/s
Best Throughput
11/11
Active Endpoints
Available via: Novita, DeepInfra, Nebius, Groq, SiliconFlow, Friendli, Cerebras, SambaNova +2 more

Leaderboard Categories