Llama 3.3 70B Instruct is Meta's open-weight model that delivers performance close to Llama 3.1 405B at a fraction of the size and cost, with a 128K context window. It is well suited for general chat, reasoning, and coding in self-hosted deployments.
Cost rate
0.2x
Context
131K
Released
Dec 6, 2024
Input
Text
Output
Text
Support
ReasoningTool callingStructured outputs
Performance
Median latency and throughput measured across recent requests.
Throughput69 tok/s
Latency2.35s
Best at
The categories where this model ranks highest.
Legal & Government
#242 路 top 61%
Medicine & Healthcare
#246 路 top 63%
Life, Physical, & Social Science
#257 路 top 61%
Entertainment, Sports, & Media
#262 路 top 62%
| Domain | Score | Rank |
|---|---|---|
| Software & IT Services | 1338 | #282 |
| Writing, Literature, & Language | 1292 | #278 |
| Life, Physical, & Social Science | 1341 | #257 |
| Entertainment, Sports, & Media | 1285 | #262 |
| Business, Management, & Finance | 1304 | #275 |
| Mathematical | 1302 | #264 |
| Legal & Government | 1342 | #242 |
| Medicine & Healthcare | 1343 | #246 |
Latest news about Llama 3.3 70B Instruct
Recent articles and community posts, linked to their original sources.

- How Meta's Llama strategy aims to become the 'Linux' of AIRick Orford - Trading Stocks and Options For All on MSN
Meta Llama for OCR is Better Than ExpectedTech Expert Tutorials1,277- Meta's Llama open-source strategy and MTIA chipsRick Orford - Trading Stocks and Options For All on MSN
How Free AI Models Like Llama 4 Are Changing the GameHarper Carroll AI8,375