Home / Models

Calculated memory fit

AI models

Each page answers how much GPU memory a model needs, and which catalog GPUs can hold it at a given quantization and context length. Figures are estimates from official model configs, not measured benchmarks.

DeepSeek

Model	Publisher	Parameters	Native context	Architecture
DeepSeek R1 Distill Qwen 14B	DeepSeek	15B	131,072	qwen2
DeepSeek R1 Distill Qwen 32B	DeepSeek	33B	131,072	qwen2
DeepSeek LLM 7B Chat	DeepSeek	6.9B	4,096	llama
DeepSeek R1 Distill Llama 70B	DeepSeek	71B	131,072	llama
DeepSeek R1 Distill Qwen 7B	DeepSeek	7.6B	131,072	qwen2
DeepSeek R1 Distill Llama 8B	DeepSeek	8.0B	131,072	llama

Falcon

Model	Publisher	Parameters	Native context	Architecture
Falcon 3 10B Instruct	TII	10B	32,768	llama
Falcon 7B Instruct	TII	7.2B	—	falcon
Falcon 3 7B Instruct	TII	7.5B	32,768	llama

Gemma

Model	Publisher	Parameters	Native context	Architecture
Gemma 2 2B Instruct	Google	2.6B	8,192	gemma2
Gemma 2 27B Instruct	Google	27B	8,192	gemma2
Gemma 2 9B Instruct	Google	9.2B	8,192	gemma2

GLM

Model	Publisher	Parameters	Native context	Architecture
GLM-4-9B-0414	Zhipu AI	9.4B	32,768	glm4
GLM-4 9B Chat	Zhipu AI	9.4B	—	chatglm

GPT-OSS

Model	Publisher	Parameters	Native context	Architecture
GPT-OSS 120B	OpenAI	117B	131,072	gpt_oss
GPT-OSS 20B	OpenAI	22B	131,072	gpt_oss

Hermes

Model	Publisher	Parameters	Native context	Architecture
Hermes 3 Llama 3.1 70B	NousResearch	71B	131,072	llama
Hermes 3 Llama 3.1 8B	NousResearch	8.0B	131,072	llama

Llama

Model	Publisher	Parameters	Native context	Architecture
Llama 3.2 1B Instruct	Meta	1.2B	131,072	llama
Code Llama 13B Instruct	Meta	13B	16,384	llama
Llama 3.2 3B Instruct	Meta	3.2B	131,072	llama
Code Llama 34B Instruct	Meta	34B	16,384	llama
Code Llama 7B Instruct	Meta	6.7B	16,384	llama
Llama 3.1 70B Instruct	Meta	71B	131,072	llama
Llama 3.3 70B Instruct	Meta	71B	131,072	llama
Llama 3.1 8B Instruct	Meta	8.0B	131,072	llama
Llama 3 8B Instruct	Meta	8.0B	8,192	llama

Mistral

Model	Publisher	Parameters	Native context	Architecture
Mistral Nemo 12B Instruct	Mistral	12B	131,072	mistral
Mistral Small 24B Instruct	Mistral	24B	32,768	mistral
Mistral 7B Instruct v0.3	Mistral	7.3B	32,768	mistral

Mixtral

Model	Publisher	Parameters	Native context	Architecture
Mixtral 8x22B Instruct	Mistral	141B	65,536	mixtral
Mixtral 8x7B Instruct	Mistral	47B	32,768	mixtral

OLMo

Model	Publisher	Parameters	Native context	Architecture
OLMo 2 13B Instruct	AllenAI	14B	4,096	olmo2
OLMo 2 7B Instruct	AllenAI	7.3B	4,096	olmo2

OpenChat

Model	Publisher	Parameters	Native context	Architecture
OpenChat 3.5 7B	OpenChat	7.2B	8,192	mistral

Phi

Model	Publisher	Parameters	Native context	Architecture
Phi-3 Medium 128K Instruct	Microsoft	14B	131,072	phi3
Phi-3 Medium 4K Instruct	Microsoft	14B	4,096	phi3
Phi-4	Microsoft	15B	16,384	phi3
Phi-3.5 Mini Instruct	Microsoft	3.8B	131,072	phi3
Phi-3 Mini 128K Instruct	Microsoft	3.8B	131,072	phi3
Phi-3 Mini 4K Instruct	Microsoft	3.8B	4,096	phi3
Phi-4 Mini Instruct	Microsoft	3.8B	131,072	phi3

Qwen

Model	Publisher	Parameters	Native context	Architecture
Qwen2.5 14B Instruct	Alibaba	15B	32,768	qwen2
Qwen3 14B	Alibaba	15B	40,960	qwen3
Qwen2.5 1.5B Instruct	Qwen	1.5B	32,768	qwen2
Qwen2.5-Coder 1.5B Instruct	Qwen	1.5B	32,768	qwen2
Qwen2.5 3B Instruct	Alibaba	3.1B	32,768	qwen2
Qwen2.5 32B Instruct	Alibaba	33B	32,768	qwen2
Qwen2.5-Coder 32B Instruct	Alibaba	33B	32,768	qwen2
QwQ 32B	Alibaba	33B	40,960	qwen2
Qwen3 32B	Alibaba	33B	40,960	qwen3
Qwen2.5 0.5B Instruct	Qwen	494M	32,768	qwen2
Qwen2.5 72B Instruct	Alibaba	73B	32,768	qwen2
Qwen2.5 7B Instruct	Alibaba	7.6B	32,768	qwen2
Qwen2.5-Coder 7B Instruct	Alibaba	7.6B	32,768	qwen2
Qwen2.5-Math 7B Instruct	Qwen	7.6B	4,096	qwen2
Qwen2 7B Instruct	Qwen	7.6B	32,768	qwen2
Qwen3 8B	Alibaba	8.2B	40,960	qwen3

SmolLM

Model	Publisher	Parameters	Native context	Architecture
SmolLM2 1.7B Instruct	HuggingFaceTB	1.7B	8,192	llama

Solar

Model	Publisher	Parameters	Native context	Architecture
SOLAR 10.7B Instruct	Upstage	11B	4,096	llama

StarCoder

Model	Publisher	Parameters	Native context	Architecture
StarCoder2 15B	BigCode	16B	16,384	starcoder2
StarCoder2 7B	BigCode	7.2B	16,384	starcoder2

TinyLlama

Model	Publisher	Parameters	Native context	Architecture
TinyLlama 1.1B Chat v1.0	TinyLlama	1.1B	2,048	llama

Vicuna

Model	Publisher	Parameters	Native context	Architecture
Vicuna 13B v1.5	LMSYS	13B	4,096	llama
Vicuna 7B v1.5	LMSYS	6.7B	4,096	llama

Yi

Model	Publisher	Parameters	Native context	Architecture
Yi 1.5 34B Chat	01.AI	34B	4,096	llama
Yi 1.5 9B Chat	01.AI	8.8B	4,096	llama