Skip to content

EuroLLM 1.7B Instruct VRAM requirements

How much VRAM EuroLLM 1.7B Instruct needs
4-bit weights: 958 MBWith 4K context: 2.1 GB8-bit, 4K context: 2.8 GB
Parameters
1.7B
Longest context
4K
Released
Aug 2024
License
Apache 2.0
Strongest at
Not rated

How much VRAM EuroLLM 1.7B Instruct needs

Memory EuroLLM 1.7B Instruct needs at each precision and context length
PrecisionWeightsWith 4K
16-bit3.1 GB4.2 GB
8-bit1.6 GB2.8 GB
6-bit1.3 GB2.4 GB
5-bit1.1 GB2.2 GB
4-bit958 MB2.1 GB
3-bit772 MB1.9 GB
2-bit593 MB1.7 GB
1.6-bit326 MB1.4 GB

One conversation, 16-bit context cache, runtime overhead included.

GPUs and Macs that run EuroLLM 1.7B Instruct

The best way EuroLLM 1.7B Instruct runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40608-bit4K118 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on RTX 4060 report
RTX 50608-bit4K184 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5060 report
RTX 30708-bit4K184 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 3070 report
RTX 3080 10 GB8-bit4K285 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 3080 10 GB report
RTX 3060 12 GB8-bit4K152 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on RTX 3060 12 GB report
Arc B5808-bit4K126 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Arc B580 report
RTX 40708-bit4K203 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4070 report
RTX 4070 Super8-bit4K203 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4070 Super report
RTX 50708-bit4K258 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5070 report
RTX 4060 Ti 16 GB8-bit4K124 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit4K184 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit4K149 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Arc A770 16 GB report
Radeon RX 9070 XT8-bit4K198 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit4K258 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4070 Ti Super report
RTX 4080 Super8-bit4K278 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4080 Super report
RTX 5070 Ti8-bit4K324 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5070 Ti report
RTX 50808-bit4K342 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5080 report
Radeon RX 7900 XTX8-bit4K265 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Radeon RX 7900 XTX report
RTX 30908-bit4K335 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 3090 report
RTX 40908-bit4K355 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 4090 report
Radeon AI PRO R97008-bit4K198 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Radeon AI PRO R9700 report
RTX 50908-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 5090 report
2× RTX 30908-bit4K335 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 2× RTX 3090 report
2× RTX 40908-bit4K355 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit4K436 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit4K342 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX 6000 Ada report
L40S8-bit4K315 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on L40S report
2× RTX 50908-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 2× RTX 5090 report
4× RTX 30908-bit4K335 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB8-bit4K59 tok/s1.7 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit4K59 tok/s1.7 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit4K59 tok/s1.7 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit4K117 tok/s0.8 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit4K117 tok/s0.8 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit4K117 tok/s0.8 sPDF: EuroLLM 1.7B Instruct on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit4K181 tok/s0.4 sPDF: EuroLLM 1.7B Instruct on M4 Max 128 GB report
M5 Max 128 GB8-bit4K195 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on M5 Max 128 GB report
M3 Ultra 96 GB8-bit4K213 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit4K213 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit4K213 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit4K262 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit4K92 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit4K92 tok/s0.2 sPDF: EuroLLM 1.7B Instruct on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit4K129 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on RTX Spark report
DGX Spark8-bit4K118 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on DGX Spark report
Dual DGX Spark8-bit4K118 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on Dual DGX Spark report
DGX Station GB3008-bit4K982 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit4K526 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit4K569 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on A100 80 GB SXM report
H100 SXM8-bit4K740 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on H100 SXM report
4× H200 NVL8-bit4K861 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 4× H200 NVL report
8× H200 NVL8-bit4K861 tok/s0.1 sPDF: EuroLLM 1.7B Instruct on 8× H200 NVL report

Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.

How to run it

llama.cpp
llama-server -hf mradermacher/EuroLLM-1.7B-Instruct-GGUF
vLLM
vllm serve utter-project/EuroLLM-1.7B-Instruct

Questions

How much VRAM does EuroLLM 1.7B Instruct need?

EuroLLM 1.7B Instruct needs 958 MB for its weights at 4-bit, 2.1 GB with an 4K context and 2.1 GB with 4K. At 8-bit, 2.8 GB.

What is the minimum VRAM for EuroLLM 1.7B Instruct?

About 1.9 GB at 3-bit with an 4K context, the lowest precision we recommend for a model this size. The smallest machine here that runs it without spilling is the RTX 4060.

How big is EuroLLM 1.7B Instruct?

1.7B parameters: 3.1 GB of weights at 16-bit, 958 MB at 4-bit.

Can EuroLLM 1.7B Instruct run on an RTX 4090?

Yes. At 8-bit it holds 4K of context and writes about 355 tokens a second.

More models