Meta

Maker of the open-weight Llama family.

Meta releases the Llama family of open-weight models, available across a wide range of inference providers including AWS Bedrock, Groq, and Fireworks.

ai.meta.com

28 model routes9 providersEU + US regions

Models from Meta

Model	Context	Zero data retention	Training	Region	Provider
Llama 4 Maverick 17B 128E Instruct FP8 Meta	1M	Enterprise	No	US	DeepInfra
Llama 4 Maverick Instruct Meta	1M	Enterprise	No	US	Novita
Llama 4 Scout 17B 16E Instruct Meta	328K	Enterprise	No	US	DeepInfra
Llama Guard 4 12B Meta	164K	Enterprise	No	US	DeepInfra
Llama 4 Scout Instruct Meta	131K	Enterprise	No	US	Novita
Llama 4 Scout 17B Meta	128K	Enterprise	No	US	Groq
Llama 3.3 70B Meta	131K	Enterprise	No	EU	Nebius
Llama 3.3 70B Instruct Meta	131K	Enterprise	No	US	Novita
Llama 3.3 70B Instruct Turbo Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.3 70B Meta	128K	Zero data retention	No	EU	Berget
Llama 3.3 70B Meta	128K	Zero data retention	No	EU	Evroc
Llama 3.3 70B Meta	128K	Enterprise	No	US	Groq
Llama 3.3 70B Instruct Meta	128K	Zero data retention	No	US	Infercom
Llama 3.3 70B Instruct Meta	66K	Enterprise	No	EU	Inceptron
Llama 3.2 11B Vision Instruct Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.2 1B Instruct Meta	131K	Enterprise	No	US	Novita
Llama 3.2 3B Instruct Meta	33K	Enterprise	No	US	Novita
Llama 3.1 70B Instruct Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.1 70B Instruct Turbo Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.1 8B Instruct Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.1 8B Instruct Turbo Meta	131K	Enterprise	No	US	DeepInfra
Llama 3.1 Nemotron Ultra 253B v1 Meta	131K	Enterprise	No	EU	Nebius
Llama 3.1 8B Meta	128K	Enterprise	No	US	Cerebras
Llama 3.1 8B Instant Meta	128K	Enterprise	No	US	Groq
Llama 3.1 8B Instruct Meta	16K	Enterprise	No	US	Novita
Llama 3 8B Instruct Meta	8K	Enterprise	No	US	Novita
Llama3 70B Instruct Meta	8K	Enterprise	No	US	Novita
Llama-3.1-8B-Instruct Meta	—	Zero data retention	No	EU	Berget