Skip to content

OLMo 2 0425 1B VRAM requirements

How much VRAM OLMo 2 0425 1B needs
4-bit weights: 859 MBWith 4K context: 2.0 GB8-bit, 4K context: 2.7 GB
Parameters
1.5B
Longest context
4K
Released
Apr 2025
License
Apache 2.0
Strongest at
Not rated

How much VRAM OLMo 2 0425 1B needs

Memory OLMo 2 0425 1B needs at each precision and context length
PrecisionWeightsWith 4K
16-bit2.8 GB4.0 GB
8-bit1.5 GB2.7 GB
6-bit1.1 GB2.3 GB
5-bit1.0 GB2.2 GB
4-bit859 MB2.0 GB
3-bit692 MB1.9 GB
2-bit531 MB1.7 GB
1.6-bit292 MB1.5 GB

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

GPUs and Macs that run OLMo 2 0425 1B

The best way OLMo 2 0425 1B runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40608-bit4K131 tok/s0.2 sPDF: OLMo 2 0425 1B on RTX 4060 report
RTX 50608-bit4K208 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5060 report
RTX 30708-bit4K208 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 3070 report
RTX 3080 10 GB8-bit4K329 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 3080 10 GB report
RTX 3060 12 GB8-bit4K170 tok/s0.2 sPDF: OLMo 2 0425 1B on RTX 3060 12 GB report
Arc B5808-bit4K144 tok/s0.1 sPDF: OLMo 2 0425 1B on Arc B580 report
RTX 40708-bit4K231 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4070 report
RTX 4070 Super8-bit4K231 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4070 Super report
RTX 50708-bit4K297 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5070 report
RTX 4060 Ti 16 GB8-bit4K138 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit4K208 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit4K172 tok/s0.1 sPDF: OLMo 2 0425 1B on Arc A770 16 GB report
Radeon RX 9070 XT8-bit4K230 tok/s0.1 sPDF: OLMo 2 0425 1B on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit4K297 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4070 Ti Super report
RTX 4080 Super8-bit4K321 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4080 Super report
RTX 5070 Ti8-bit4K378 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5070 Ti report
RTX 50808-bit4K400 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5080 report
Radeon RX 7900 XTX8-bit4K316 tok/s0.1 sPDF: OLMo 2 0425 1B on Radeon RX 7900 XTX report
RTX 30908-bit4K391 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 3090 report
RTX 40908-bit4K416 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 4090 report
Radeon AI PRO R97008-bit4K230 tok/s0.1 sPDF: OLMo 2 0425 1B on Radeon AI PRO R9700 report
RTX 50908-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 5090 report
2× RTX 30908-bit4K391 tok/s0.1 sPDF: OLMo 2 0425 1B on 2× RTX 3090 report
2× RTX 40908-bit4K416 tok/s0.1 sPDF: OLMo 2 0425 1B on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit4K520 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit4K400 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX 6000 Ada report
L40S8-bit4K367 tok/s0.1 sPDF: OLMo 2 0425 1B on L40S report
2× RTX 50908-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on 2× RTX 5090 report
4× RTX 30908-bit4K391 tok/s0.1 sPDF: OLMo 2 0425 1B on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB8-bit4K66 tok/s1.6 sPDF: OLMo 2 0425 1B on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit4K66 tok/s1.6 sPDF: OLMo 2 0425 1B on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit4K66 tok/s1.6 sPDF: OLMo 2 0425 1B on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit4K136 tok/s0.8 sPDF: OLMo 2 0425 1B on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit4K136 tok/s0.8 sPDF: OLMo 2 0425 1B on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit4K136 tok/s0.8 sPDF: OLMo 2 0425 1B on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit4K217 tok/s0.4 sPDF: OLMo 2 0425 1B on M4 Max 128 GB report
M5 Max 128 GB8-bit4K236 tok/s0.1 sPDF: OLMo 2 0425 1B on M5 Max 128 GB report
M3 Ultra 96 GB8-bit4K261 tok/s0.2 sPDF: OLMo 2 0425 1B on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit4K261 tok/s0.2 sPDF: OLMo 2 0425 1B on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit4K261 tok/s0.2 sPDF: OLMo 2 0425 1B on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit4K331 tok/s0.1 sPDF: OLMo 2 0425 1B on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit4K104 tok/s0.2 sPDF: OLMo 2 0425 1B on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit4K104 tok/s0.2 sPDF: OLMo 2 0425 1B on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit4K144 tok/s0.1 sPDF: OLMo 2 0425 1B on RTX Spark report
DGX Spark8-bit4K132 tok/s0.1 sPDF: OLMo 2 0425 1B on DGX Spark report
Dual DGX Spark8-bit4K132 tok/s0.1 sPDF: OLMo 2 0425 1B on Dual DGX Spark report
DGX Station GB3008-bit4K1,327 tok/s0.1 sPDF: OLMo 2 0425 1B on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit4K640 tok/s0.1 sPDF: OLMo 2 0425 1B on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit4K698 tok/s0.1 sPDF: OLMo 2 0425 1B on A100 80 GB SXM report
H100 SXM8-bit4K944 tok/s0.1 sPDF: OLMo 2 0425 1B on H100 SXM report
4× H200 NVL8-bit4K1,130 tok/s0.1 sPDF: OLMo 2 0425 1B on 4× H200 NVL report
8× H200 NVL8-bit4K1,130 tok/s0.1 sPDF: OLMo 2 0425 1B 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 allenai/OLMo-2-0425-1B-GGUF
vLLM
vllm serve allenai/OLMo-2-0425-1B

Questions

How much VRAM does OLMo 2 0425 1B need?

OLMo 2 0425 1B needs 859 MB for its weights at 4-bit, 2.0 GB with an 4K context and 2.0 GB with 4K. At 8-bit, 2.7 GB.

What is the minimum VRAM for OLMo 2 0425 1B?

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 OLMo 2 0425 1B?

1.5B parameters: 2.8 GB of weights at 16-bit, 859 MB at 4-bit.

Can OLMo 2 0425 1B run on an RTX 4090?

Yes. At 8-bit it holds 4K of context and writes about 416 tokens a second.

More models