Skip to content

OLMo 2 1124 7B Instruct VRAM requirements

How much VRAM OLMo 2 1124 7B Instruct needs
4-bit weights: 4.1 GBWith 4K context: 6.9 GB8-bit, 4K context: 10.0 GB
Parameters
7.3B
Longest context
4K
Released
Dec 2024
License
Apache 2.0
Strongest at
Not rated

How much VRAM OLMo 2 1124 7B Instruct needs

Memory OLMo 2 1124 7B Instruct needs at each precision and context length
PrecisionWeightsWith 4K
16-bit13.6 GB16.4 GB
8-bit7.2 GB10.0 GB
6-bit5.6 GB8.3 GB
5-bit4.8 GB7.6 GB
4-bit4.1 GB6.9 GB
3-bit3.3 GB6.1 GB
2-bit2.5 GB5.3 GB
1.6-bit1.4 GB4.2 GB

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

GPUs and Macs that run OLMo 2 1124 7B Instruct

The best way OLMo 2 1124 7B Instruct runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40604-bit4K44 tok/s1.0 sPDF: OLMo 2 1124 7B Instruct on RTX 4060 report
RTX 50604-bit4K70 tok/s0.8 sPDF: OLMo 2 1124 7B Instruct on RTX 5060 report
RTX 30704-bit4K70 tok/s0.7 sPDF: OLMo 2 1124 7B Instruct on RTX 3070 report
RTX 3080 10 GB6-bit4K86 tok/s0.5 sPDF: OLMo 2 1124 7B Instruct on RTX 3080 10 GB report
RTX 3060 12 GB8-bit4K33 tok/s1.1 sPDF: OLMo 2 1124 7B Instruct on RTX 3060 12 GB report
Arc B5808-bit4K29 tok/s0.8 sPDF: OLMo 2 1124 7B Instruct on Arc B580 report
RTX 40708-bit4K46 tok/s0.5 sPDF: OLMo 2 1124 7B Instruct on RTX 4070 report
RTX 4070 Super8-bit4K46 tok/s0.4 sPDF: OLMo 2 1124 7B Instruct on RTX 4070 Super report
RTX 50708-bit4K61 tok/s0.5 sPDF: OLMo 2 1124 7B Instruct on RTX 5070 report
RTX 4060 Ti 16 GB8-bit4K27 tok/s0.7 sPDF: OLMo 2 1124 7B Instruct on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit4K41 tok/s0.6 sPDF: OLMo 2 1124 7B Instruct on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit4K36 tok/s0.7 sPDF: OLMo 2 1124 7B Instruct on Arc A770 16 GB report
Radeon RX 9070 XT8-bit4K48 tok/s0.4 sPDF: OLMo 2 1124 7B Instruct on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit4K61 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on RTX 4070 Ti Super report
RTX 4080 Super8-bit4K66 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on RTX 4080 Super report
RTX 5070 Ti8-bit4K79 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on RTX 5070 Ti report
RTX 50808-bit4K84 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on RTX 5080 report
Radeon RX 7900 XTX8-bit4K70 tok/s0.6 sPDF: OLMo 2 1124 7B Instruct on Radeon RX 7900 XTX report
RTX 30908-bit4K82 tok/s0.4 sPDF: OLMo 2 1124 7B Instruct on RTX 3090 report
RTX 40908-bit4K88 tok/s0.2 sPDF: OLMo 2 1124 7B Instruct on RTX 4090 report
Radeon AI PRO R97008-bit4K48 tok/s0.4 sPDF: OLMo 2 1124 7B Instruct on Radeon AI PRO R9700 report
RTX 50908-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on RTX 5090 report
2× RTX 30908-bit4K82 tok/s0.2 sPDF: OLMo 2 1124 7B Instruct on 2× RTX 3090 report
2× RTX 40908-bit4K88 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit4K114 tok/s0.2 sPDF: OLMo 2 1124 7B Instruct on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit4K84 tok/s0.2 sPDF: OLMo 2 1124 7B Instruct on RTX 6000 Ada report
L40S8-bit4K77 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on L40S report
2× RTX 50908-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 2× RTX 5090 report
4× RTX 30908-bit4K82 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB8-bit4K13 tok/s9.2 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit4K13 tok/s9.2 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit4K13 tok/s9.2 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit4K28 tok/s4.6 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit4K28 tok/s4.6 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit4K28 tok/s4.6 sPDF: OLMo 2 1124 7B Instruct on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit4K49 tok/s2.3 sPDF: OLMo 2 1124 7B Instruct on M4 Max 128 GB report
M5 Max 128 GB8-bit4K54 tok/s0.7 sPDF: OLMo 2 1124 7B Instruct on M5 Max 128 GB report
M3 Ultra 96 GB8-bit4K61 tok/s1.3 sPDF: OLMo 2 1124 7B Instruct on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit4K61 tok/s1.3 sPDF: OLMo 2 1124 7B Instruct on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit4K61 tok/s1.3 sPDF: OLMo 2 1124 7B Instruct on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit4K84 tok/s0.4 sPDF: OLMo 2 1124 7B Instruct on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit4K20 tok/s1.3 sPDF: OLMo 2 1124 7B Instruct on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit4K20 tok/s1.3 sPDF: OLMo 2 1124 7B Instruct on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit4K28 tok/s0.5 sPDF: OLMo 2 1124 7B Instruct on RTX Spark report
DGX Spark8-bit4K25 tok/s0.5 sPDF: OLMo 2 1124 7B Instruct on DGX Spark report
Dual DGX Spark8-bit4K25 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on Dual DGX Spark report
DGX Station GB3008-bit4K411 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit4K147 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit4K164 tok/s0.3 sPDF: OLMo 2 1124 7B Instruct on A100 80 GB SXM report
H100 SXM8-bit4K245 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on H100 SXM report
4× H200 NVL8-bit4K319 tok/s0.1 sPDF: OLMo 2 1124 7B Instruct on 4× H200 NVL report
8× H200 NVL8-bit4K319 tok/s0.1 sPDF: OLMo 2 1124 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 bartowski/OLMo-2-1124-7B-Instruct-GGUF
vLLM
vllm serve allenai/OLMo-2-1124-7B-Instruct
MLX (Mac)
mlx_lm.chat --model mlx-community/OLMo-2-1124-7B-Instruct-4bit

Questions

How much VRAM does OLMo 2 1124 7B Instruct need?

OLMo 2 1124 7B Instruct needs 4.1 GB for its weights at 4-bit, 6.9 GB with an 4K context and 6.9 GB with 4K. At 8-bit, 10.0 GB.

What is the minimum VRAM for OLMo 2 1124 7B Instruct?

About 6.1 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 1124 7B Instruct?

7.3B parameters: 13.6 GB of weights at 16-bit, 4.1 GB at 4-bit.

Can OLMo 2 1124 7B Instruct run on an RTX 4090?

Yes. At 8-bit it holds 4K of context and writes about 88 tokens a second.

More models