Llama 3 8B
An 8-billion-parameter Meta Llama 3 auto-regressive language model, pretrained and instruction-tuned, intended for commercial and research use in English, with instruction-tuned versions built for assistant-like chat and pretrained ones for natural language generation.
text → text + code · made by Meta
$0.3→$0.6per Mtok in / out
AWS Bedrock
Sold by 1 way
| Seller | Lane | Rate |
|---|---|---|
| AWS Bedrock 0 ways | $0.3per Mtok in$0.6per Mtok out | |
| AWS Bedrock cloudnot on the seller's list today | standard | $0.3per Mtok in$0.6per Mtok outpricing.us-east-1.amazonaws.com · read 2026-08-24 |
Measured 3 standings
| Place | Board | Metric | Score |
|---|---|---|---|
| 49thof 137 | Wino grande — Epoch AI | Accuracy | 75.7 |
| 79thof 81 | Forecastbench — Epoch AI | Overall score | 52.9 |
| 601stof 656 | Epoch capabilities index — Epoch AI | ECI Score | 116.33 |
About
Llama 3 8B — a text and code model from Meta, placed 49th of 137 on Wino grande — Epoch AI.
It takes text and returns text and code, with a context window of 8,192 tokens. The catalogue files it under code and agents. It has been measured on 3 boards, and stands best at 49th of 137 on Wino grande — Epoch AI.
Every current figure
- Maker
- Meta
- Register
- model
- Takes
- text
- Returns
- text + code
- Context
- 8,192 tokens
- Parameters
- 8 billion · read from its own name
- Boards
- 3
- Best place
- 49th of 137 — Wino grande — Epoch AI