Llama Models

Q: Which Llama model should I start with?

llama-4-maverick at $0.2/M input — the cheapest entry here that declares tool calling, and it carries a 1.05M context. Move up to aihubmix-Llama-3-1-405B-Instruct when answer quality matters more than cost.

42 modelsGeneral models from $0.2/M inputUp to 1.05M context

All 42 Llama Models

Open in model list

Llama models on AIHubMix with input and output modalities, context length, maximum output, price per million tokens including cache read and cache write rates, and measured throughput and latency.
	Modalities
llama-4-maverick	Takes text, vision, returns text.	1.05M	$0.2$0.2/M	98 tok/s	0.23 s
llama-4-scout	Takes text, vision, returns text.	131K	$0.2$0.2/M	148 tok/s	0.89 s
llama-3.3-70b		131K	$0.6$0.6/M	3250 tok/s	0.21 s
llama3-groq-8b-8192-tool-use-preview		—	$0.00019$0.00019/M	—	—
llama3-groq-70b-8192-tool-use-preview		—	$0.00089$0.00089/M	—	—
meta-llama/llama-3.1-405b-instruct:free		—	$0.02$0.02/M	—	—
meta-llama/llama-3.1-70b-instruct:free		—	$0.02$0.02/M	—	—
meta-llama/llama-3.1-8b-instruct:free		—	$0.02$0.02/M	—	—
meta-llama/llama-3.2-11b-vision-instruct:free		—	$0.02$0.02/M	—	—
meta-llama/llama-3.2-3b-instruct:free		—	$0.02$0.02/M	—	—
llama3-8b-8192		—	$0.06$0.12/M	—	—
llama2-7b-2048		—	$0.1$0.1/M	—	—
deepseek-r1-distill-qianfan-llama-8b		—	$0.137$0.548/M	—	—
llama-3.2-11b-vision-preview		—	$0.2$0.2/M	—	—
llama-3.2-1b-preview		—	$0.2$0.2/M	—	—
llama-3.2-3b-preview		—	$0.2$0.2/M	—	—
qianfan-llama-vl-8b		—	$0.274$0.685/M	—	—
aihubmix-Llama-3-1-8B-Instruct		—	$0.3$0.6/M	—	—
llama-3.1-8b-instant		—	$0.3$0.6/M	—	—
llama3-8b-8192(33)		—	$0.3$0.3/M	—	—
llama3.1-8b		—	$0.3$0.6/M	—	—
meta/llama3-8B-chat		—	$0.3$0.3/M	—	—
aihubmix-Llama-3-2-11B-Vision		—	$0.4$0.4/M	—	—
Gryphe/MythoMax-L2-13b		—	$0.4$0.4/M	—	—
llama2-70b-4096	Takes , returns text.	—	$0.5$0.5/M	—	—
meta-llama/Llama-3.2-90B-Vision-Instruct		—	$0.5$0.5/M	—	—
meta-llama-3-8b		—	$0.548$0.548/M	—	—
aihubmix-Llama-3-1-70B-Instruct		—	$0.6$0.78/M	—	—
llama-3.1-70b		—	$0.6$0.6/M	—	—
llama-3.1-70b-versatile		—	$0.6$0.6/M	—	—
aihubmix-Llama-3-70B-Instruct		—	$0.7$0.7/M	—	—
llama3-70b-8192		—	$0.7$0.9373/M	—	—
qianfan-chinese-llama-2-13b		—	$0.822$0.822/M	—	—
WizardLM/WizardCoder-Python-34B-V1.0		—	$0.9$0.9/M	—	—
aihubmix-Llama-3-2-90B-Vision		—	$2.4$2.4/M	—	—
llama-3.2-90b-vision-preview		—	$2.4$2.4/M	—	—
llama3-70b-8192(33)		—	$2.65$2.65/M	—	—
llama-3.1-405b-instruct		—	$4$4/M	—	—
llama-3.1-405b-reasoning		—	$4$4/M	—	—
meta-llama-3-70b		—	$4.795$4.795/M	—	—
aihubmix-Llama-3-1-405B-Instruct		—	$5$15/M	—	—
meta/llama-3.1-405b-instruct		—	$5$5/M	—	—

Prices are USD per million tokens; cache read and cache write are the rates for prompt-cache hits and for writing a prompt into the cache. Throughput and latency are measured on AIHubMix — the same figures the model detail page shows — not vendor claims. A dash means the catalog does not publish that field for that model, which is not the same as the model not supporting it.

Common Comparisons

llama-4-maverick vs llama-4-scoutllama-4-maverick vs llama2-70b-4096gpt-5.6-luna vs llama-4-maverickdeepseek-v4-flash vs llama-4-maverickahm-Phi-3-5-vision-instruct vs llama-4-maverick

Llama on AIHubMix

Which Llama model should I start with?

llama-4-maverick at $0.2/M input — the cheapest entry here that declares tool calling, and it carries a 1.05M context. Move up to aihubmix-Llama-3-1-405B-Instruct when answer quality matters more than cost.

Why are there several entries for the same model?

Because each row is a route you can call, not a model release. Some IDs name an upstream (azure-, alicloud-, cc-), some are the open-weight repository form (meta-llama/…), and some differ only in capitalisation, kept so older integrations keep working.

The catalog does not carry a field saying which of those a given row is, so this page does not sort them into buckets it would have to invent. Every row shows that route’s own price, context and speed — compare those directly, and open a model to see the upstreams that serve it.

Do I need a separate Llama account?

No. One AIHubMix key covers every model on this page, and switching between them is a change to the model string — billing, rate limits, and logs stay in one place.