Skip to content

D1 Omni 600M VRAM requirements

How much VRAM D1 Omni 600M needs
4-bit weights: 339 MBWith 8K context: 1.0 GBWith 32K context: 1.3 GB8-bit, 8K context: 1.3 GB
Parameters
587M
Longest context
125K
Released
Oct 2026
License
Lfm1.0
Strongest at
Text and images

Its licence (Lfm1.0) is not a standard open-source one: check the terms before commercial use.

How much VRAM D1 Omni 600M needs

Memory D1 Omni 600M needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 125K
16-bit1.1 GB1.8 GB2.1 GB3.2 GB
8-bit595 MB1.3 GB1.6 GB2.7 GB
6-bit459 MB1.1 GB1.4 GB2.5 GB
5-bit398 MB1.1 GB1.4 GB2.5 GB
4-bit339 MB1.0 GB1.3 GB2.4 GB
3-bit274 MB1.0 GB1.2 GB2.3 GB
2-bit210 MB927 MB1.2 GB2.3 GB
1.6-bit115 MB832 MB1.1 GB2.2 GB

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

GPUs and Macs that run D1 Omni 600M

The best way D1 Omni 600M runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40608-bit125K296 tok/s0.1 sPDF: D1 Omni 600M on RTX 4060 report
RTX 50608-bit125K446 tok/s0.1 sPDF: D1 Omni 600M on RTX 5060 report
RTX 30708-bit125K446 tok/s0.1 sPDF: D1 Omni 600M on RTX 3070 report
RTX 3080 10 GB8-bit125K658 tok/s0.1 sPDF: D1 Omni 600M on RTX 3080 10 GB report
RTX 3060 12 GB8-bit125K374 tok/s0.1 sPDF: D1 Omni 600M on RTX 3060 12 GB report
Arc B5808-bit125K296 tok/s0.1 sPDF: D1 Omni 600M on Arc B580 report
RTX 40708-bit125K489 tok/s0.1 sPDF: D1 Omni 600M on RTX 4070 report
RTX 4070 Super8-bit125K489 tok/s0.1 sPDF: D1 Omni 600M on RTX 4070 Super report
RTX 50708-bit125K604 tok/s0.1 sPDF: D1 Omni 600M on RTX 5070 report
RTX 4060 Ti 16 GB8-bit125K310 tok/s0.1 sPDF: D1 Omni 600M on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit125K446 tok/s0.1 sPDF: D1 Omni 600M on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit125K341 tok/s0.1 sPDF: D1 Omni 600M on Arc A770 16 GB report
Radeon RX 9070 XT8-bit125K447 tok/s0.1 sPDF: D1 Omni 600M on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit125K604 tok/s0.1 sPDF: D1 Omni 600M on RTX 4070 Ti Super report
RTX 4080 Super8-bit125K644 tok/s0.1 sPDF: D1 Omni 600M on RTX 4080 Super report
RTX 5070 Ti8-bit125K735 tok/s0.1 sPDF: D1 Omni 600M on RTX 5070 Ti report
RTX 50808-bit125K768 tok/s0.1 sPDF: D1 Omni 600M on RTX 5080 report
Radeon RX 7900 XTX8-bit125K568 tok/s0.1 sPDF: D1 Omni 600M on Radeon RX 7900 XTX report
RTX 30908-bit125K756 tok/s0.1 sPDF: D1 Omni 600M on RTX 3090 report
RTX 40908-bit125K792 tok/s0.1 sPDF: D1 Omni 600M on RTX 4090 report
Radeon AI PRO R97008-bit125K447 tok/s0.1 sPDF: D1 Omni 600M on Radeon AI PRO R9700 report
RTX 50908-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on RTX 5090 report
2× RTX 30908-bit125K756 tok/s0.1 sPDF: D1 Omni 600M on 2× RTX 3090 report
2× RTX 40908-bit125K792 tok/s0.1 sPDF: D1 Omni 600M on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit125K937 tok/s0.1 sPDF: D1 Omni 600M on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit125K768 tok/s0.1 sPDF: D1 Omni 600M on RTX 6000 Ada report
L40S8-bit125K718 tok/s0.1 sPDF: D1 Omni 600M on L40S report
2× RTX 50908-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on 2× RTX 5090 report
4× RTX 30908-bit125K756 tok/s0.1 sPDF: D1 Omni 600M on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB8-bit125K145 tok/s0.7 sPDF: D1 Omni 600M on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit125K145 tok/s0.7 sPDF: D1 Omni 600M on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit125K145 tok/s0.7 sPDF: D1 Omni 600M on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit125K267 tok/s0.3 sPDF: D1 Omni 600M on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit125K267 tok/s0.3 sPDF: D1 Omni 600M on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit125K267 tok/s0.3 sPDF: D1 Omni 600M on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit125K380 tok/s0.2 sPDF: D1 Omni 600M on M4 Max 128 GB report
M5 Max 128 GB8-bit125K403 tok/s0.1 sPDF: D1 Omni 600M on M5 Max 128 GB report
M3 Ultra 96 GB8-bit125K431 tok/s0.1 sPDF: D1 Omni 600M on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit125K431 tok/s0.1 sPDF: D1 Omni 600M on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit125K431 tok/s0.1 sPDF: D1 Omni 600M on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit125K503 tok/s0.1 sPDF: D1 Omni 600M on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit125K227 tok/s0.1 sPDF: D1 Omni 600M on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit125K227 tok/s0.1 sPDF: D1 Omni 600M on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit125K321 tok/s0.1 sPDF: D1 Omni 600M on RTX Spark report
DGX Spark8-bit125K297 tok/s0.1 sPDF: D1 Omni 600M on DGX Spark report
Dual DGX Spark8-bit125K297 tok/s0.1 sPDF: D1 Omni 600M on Dual DGX Spark report
DGX Station GB3008-bit125K1,692 tok/s0.1 sPDF: D1 Omni 600M on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit125K1,086 tok/s0.1 sPDF: D1 Omni 600M on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit125K1,153 tok/s0.1 sPDF: D1 Omni 600M on A100 80 GB SXM report
H100 SXM8-bit125K1,397 tok/s0.1 sPDF: D1 Omni 600M on H100 SXM report
4× H200 NVL8-bit125K1,552 tok/s0.1 sPDF: D1 Omni 600M on 4× H200 NVL report
8× H200 NVL8-bit125K1,552 tok/s0.1 sPDF: D1 Omni 600M 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 LiquidAI/d1-omni-600M-GGUF
vLLM
vllm serve LiquidAI/d1-omni-600M

Questions

How much VRAM does D1 Omni 600M need?

D1 Omni 600M needs 339 MB for its weights at 4-bit, 1.0 GB with an 8K context and 1.3 GB with 32K. At 8-bit, 1.3 GB.

What is the minimum VRAM for D1 Omni 600M?

About 1.0 GB at 3-bit with an 8K 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 D1 Omni 600M?

587M parameters: 1.1 GB of weights at 16-bit, 339 MB at 4-bit.

Can D1 Omni 600M run on an RTX 4090?

Yes. At 8-bit it holds 125K of context and writes about 792 tokens a second.

More models