How much GPU memory does dolphin-2.5-mixtral-8x7b need?

About 107.2 GB for conversations of up to 8,192 tokens: 93.41 GB of weights, the cache for one full-length conversation and the runtime's own memory.

What is the cheapest GPU to run dolphin-2.5-mixtral-8x7b?

Right now 8x RTX 3090 at $2.28 an hour; 4x RTX 5090 at $2.51 an hour is the best value for its speed. Prices are live offers from GPU clouds, checked every 15 minutes.

Can dolphin-2.5-mixtral-8x7b run on an RTX 4090?

Not on one: it needs about 107.2 GB and an RTX 4090 has 24 GB. Use a bigger GPU, or a smaller (quantized) version of the model.

How do I deploy dolphin-2.5-mixtral-8x7b as an API?

Pick a machine on its GridPup page and press Deploy: GridPup rents the GPU, downloads the model from Hugging Face and serves it at an https address with keys and OpenAI's chat API, billed by the hour.

Models / dolphin-2.5-mixtral-8x7b / dolphin-2.5-mixtral-8x7b ~46.7B

dphn/dolphin-2.5-mixtral-8x7b

Text generation · served with vLLM · ↓ 1.2K downloads a month · View on Hugging Face

dphn/dolphin-2.5-mixtral-8x7b needs about 107.2 GB of GPU memory for conversations of up to 8,192 tokens. It runs on 8x RTX 3090 from $2.28/h right now; best value: 4x RTX 5090 at $2.51/h.

Weights: 93 GB
GPU memory needed: about 107.2 GB
Conversation length: 8,192 tokens
Model allows: 32,768 tokens
Memory · disk: 54.7 GB · 143 GB
Architecture: MixtralForCausalLM
Licence: apache-2.0

Versions of dolphin-2.5-mixtral-8x7b ~46.7B

The same model packaged differently: a quantized version needs less GPU memory and runs on a cheaper machine.

Version	Runtime	Weights	GPU memory	Cheapest machine now
Originaldphn/dolphin-2.5-mixtral-8x7b	vLLM	93 GB	107.2 GB	$2.25/h 8x RTX 3090
GPTQ (TheBloke)TheBloke/dolphin-2.5-mixtral-8x7b-GPTQ	vLLM	24 GB	29.9 GB	$0.44/h RTX 5000 Ada

Hardware and deploy

Endpoint name

$2.51/h from your balance while it runs, billed by the hour; stop it from its page. GridPup rents the machine, which downloads the model from Hugging Face (usually ready in 10 to 15 minutes), and gives you an https address, keys and a Playground.

Questions

How much GPU memory does dolphin-2.5-mixtral-8x7b need?: About 107.2 GB for conversations of up to 8,192 tokens: 93.41 GB of weights, the cache for one full-length conversation and the runtime's own memory.
What is the cheapest GPU to run dolphin-2.5-mixtral-8x7b?: Right now 8x RTX 3090 at $2.28 an hour; 4x RTX 5090 at $2.51 an hour is the best value for its speed. Prices are live offers from GPU clouds, checked every 15 minutes.
Can dolphin-2.5-mixtral-8x7b run on an RTX 4090?: Not on one: it needs about 107.2 GB and an RTX 4090 has 24 GB. Use a bigger GPU, or a smaller (quantized) version of the model.
How do I deploy dolphin-2.5-mixtral-8x7b as an API?: Pick a machine on its GridPup page and press Deploy: GridPup rents the GPU, downloads the model from Hugging Face and serves it at an https address with keys and OpenAI's chat API, billed by the hour.

Loading…

dphn/dolphin-2.5-mixtral-8x7b

Text generation · served with vLLM · ↓ 1.2K downloads a month · View on Hugging Face

dphn/dolphin-2.5-mixtral-8x7b needs about 107.2 GB of GPU memory for conversations of up to 8,192 tokens. It runs on 8x RTX 3090 from $2.28/h right now; best value: 4x RTX 5090 at $2.51/h.

Weights

93 GB

GPU memory needed

about 107.2 GB

Conversation length

8,192 tokens

Model allows

32,768 tokens

Memory · disk

54.7 GB · 143 GB

Architecture

MixtralForCausalLM

Licence

apache-2.0

Versions of dolphin-2.5-mixtral-8x7b ~46.7B

The same model packaged differently: a quantized version needs less GPU memory and runs on a cheaper machine.

Version	Runtime	Weights	GPU memory	Cheapest machine now
Originaldphn/dolphin-2.5-mixtral-8x7b	vLLM	93 GB	107.2 GB	$2.25/h 8x RTX 3090
GPTQ (TheBloke)TheBloke/dolphin-2.5-mixtral-8x7b-GPTQ	vLLM	24 GB	29.9 GB	$0.44/h RTX 5000 Ada

Hardware and deploy

Endpoint name

Questions

How much GPU memory does dolphin-2.5-mixtral-8x7b need?: About 107.2 GB for conversations of up to 8,192 tokens: 93.41 GB of weights, the cache for one full-length conversation and the runtime's own memory.
What is the cheapest GPU to run dolphin-2.5-mixtral-8x7b?: Right now 8x RTX 3090 at $2.28 an hour; 4x RTX 5090 at $2.51 an hour is the best value for its speed. Prices are live offers from GPU clouds, checked every 15 minutes.
Can dolphin-2.5-mixtral-8x7b run on an RTX 4090?: Not on one: it needs about 107.2 GB and an RTX 4090 has 24 GB. Use a bigger GPU, or a smaller (quantized) version of the model.
How do I deploy dolphin-2.5-mixtral-8x7b as an API?: Pick a machine on its GridPup page and press Deploy: GridPup rents the GPU, downloads the model from Hugging Face and serves it at an https address with keys and OpenAI's chat API, billed by the hour.