Inference
EN
Meta

Llama 3.1 8B

Llama 3.1 8B is a standard model from Meta. 10 benchmarks count toward its index, in 5 categories.

  • Open weights
  • Standard
  • Input: text
  • Released 2024-07
RankedEnough results for an index.
IndexAll categories, out of 100.17.9±9.6Median 51.0
CoverageIndex weight that has results.55%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.128KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

8.03B parameters
16-bit
17.9 GB
8-bit
10.0 GB
4-bit
6.1 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 8 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.None counted——
CodingCode writing and repair.1 counted13.9—
ReasoningLogic problems and puzzles.None counted——
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.4 counted14.4264 / 268
MultilingualTasks in many languages.2 counted24.88 / 9
InstructionTasks with strict prompt rules.1 counted25.1—
MathMath problems.2 counted10.0167 / 168

Results

10 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
EuroEval GermanMultilingual36.6% ±0.895 / 191
EuroEval PolishMultilingual41.4% ±0.4105 / 209
EuroEval DutchMultilingual43.8% ±0.489 / 175
EuroEval SwedishMultilingual40.6% ±0.4102 / 195
EuroEval FrenchMultilingualAbove fitFar above this model's index.45.9% ±0.595 / 178
EuroEval PortugueseMultilingual41.1% ±0.5103 / 182
EuroEval ItalianMultilingual38.5% ±0.699 / 174
Multilingual Grade School MathMultilingual68.9%8 / 13
EuroEval SpanishMultilingual31.3% ±0.5110 / 176
Instruction-Following EvalInstruction80.4%54 / 70
MATH level 5Math22.9% ±0.969 / 88
Massive Multitask Language UnderstandingKnowledge73.0%29 / 34
Artificial Analysis Humanity's Last ExamKnowledge4.3%190 / 222
Evaluating Large Language Models Trained on CodeCoding72.6%19 / 22
Massive Multitask Language Understanding ProfessionalKnowledge48.3%67 / 75
OTIS Mock AIME 2024-2025Math1.7% ±0.9188 / 200
GPQA diamond · Epoch AIKnowledge27.0% ±1.9205 / 217
Graduate-Level Google-Proof Q&AKnowledge30.4%134 / 141
Artificial Analysis GPQA DiamondKnowledge27.0%196 / 199
Grade School Math 8KMath84.5%—

Sources

BenchLMCC BY-NC 4.0EuroEvalMITEpoch AICC BY

History

5 releases
Index19.1 → 17.9
23 Oct 202612 Nov 2026

Questions

How good is Llama 3.1 8B?
Its index is 17.9 out of 100. That is place 279 of 283 ranked models.
What is Llama 3.1 8B best at?
Multilingual: 24.8 out of 100, place 8 of 9.
What is the context window of Llama 3.1 8B?
128K tokens.
Is Llama 3.1 8B open source?
Yes. Its weights are published.
When was Llama 3.1 8B released?
July 2024.

Badge

Llama 3.1 8B on Inference360: place 279 of 283, index 17.9
HTML
<a href="https://inference360.ai/models/llama-3-1-8b"><img src="https://inference360.ai/badges/llama-3-1-8b.svg" alt="Llama 3.1 8B on Inference360: place 279 of 283, index 17.9" height="20"></a>
Markdown
[![Llama 3.1 8B on Inference360: place 279 of 283, index 17.9](https://inference360.ai/badges/llama-3-1-8b.svg)](https://inference360.ai/models/llama-3-1-8b)

Compare with

Granite-4.0-H-1B16.3Gemini 1.5 Flash20.2MiniCPM5-1B15.5Gemma 4 E2B21.9Llama 3.1 70B24.8

More from Meta

All models
Llama 3.1 405B27.0Llama 3.1 70B24.8Llama 3.1 70B InstructUnrankedLlama 3.1 8B InstructUnrankedLlama 3.2 1B InstructUnranked