Run Llama 3.2 90B Vision (Preview) on your data

Llama 3.2 90B Vision (Preview) is a Multimodal LLM designed for visual question answering, image captioning, and document visual question answering. It excels in general knowledge, long-form text generation, multilingual translation, coding, math, and advanced reasoning.

Some noteworthy use cases of Llama 3.2 90B Vision (Preview) include image-text retrieval, visual grounding, and visual reasoning.

Metric	Value
Parameter Count	90 billion
Mixture of Experts	Unknown
Context Length	128,000 tokens
Multilingual	Yes (text-only)
Quantized*	Unknown

*Quantization is specific to the inference provider and may vary by provider.

	Modality	Price (1M tokens)
Llama 3.1 405B Instruct	Fireworks AI	text	text	$3.00	$3.00
Llama 3.1 70B Instruct	Fireworks AI	text	text	$0.90	$0.90
Llama 3.1 8B	Groq	text	text	$0.05	$0.08
Llama 3.1 8B Instruct	Fireworks AI	text	text	$0.20	$0.20
Llama 3.2 3B Instruct	Lambda Labs	text	text	$0.02	$0.02
Llama 3.3 70B Instruct	Fireworks AI	text	text	$0.90	$0.90
Llama 3.3 70B Speculative Decoding	Groq	text	text	$0.59	$0.59
Llama 3.3 70B Versatile 128k	Groq	text	text	$0.59	$0.79
Llama 4 Maverick	Fireworks AI	text, image	text	$0.22	$0.88
Llama 4 Scout	Fireworks AI	text, image	text	$0.15	$0.60

Modality

Price (1M tokens)

Model

Inference provider

Input

Output

Input

Output

Llama 3.1 405B Instruct

Fireworks AI