Skip to content

VLX Seek 1.5 10B VRAM requirements

How much VRAM VLX Seek 1.5 10B needs
4-bit weights: 5.7 GBWith 8K context: 7.0 GBWith 32K context: 7.8 GB8-bit, 8K context: 11.3 GB
Parameters
10B
Longest context
256K
Released
Jul 2026
License
Apache 2.0
Strongest at
Text and images

How much VRAM VLX Seek 1.5 10B needs

Memory VLX Seek 1.5 10B needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
16-bit18.7 GB20.0 GB20.8 GB23.8 GB
8-bit9.9 GB11.3 GB12.0 GB15.0 GB
6-bit7.6 GB9.0 GB9.7 GB12.7 GB
5-bit6.6 GB8.0 GB8.7 GB11.7 GB
4-bit5.7 GB7.0 GB7.8 GB10.8 GB
3-bit4.6 GB5.9 GB6.7 GB9.7 GB
2-bit3.5 GB4.8 GB5.6 GB8.6 GB
1.6-bit1.9 GB3.3 GB4.0 GB7.0 GB

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

GPUs and Macs that run VLX Seek 1.5 10B

The best way VLX Seek 1.5 10B runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40604-bit24K33 tok/s1.2 sPDF: VLX Seek 1.5 10B on RTX 4060 report
RTX 50604-bit24K54 tok/s0.9 sPDF: VLX Seek 1.5 10B on RTX 5060 report
RTX 30704-bit24K54 tok/s0.9 sPDF: VLX Seek 1.5 10B on RTX 3070 report
RTX 3080 10 GB6-bit24K67 tok/s0.6 sPDF: VLX Seek 1.5 10B on RTX 3080 10 GB report
RTX 3060 12 GB6-bit80K33 tok/s1.4 sPDF: VLX Seek 1.5 10B on RTX 3060 12 GB report
Arc B5806-bit80K29 tok/s1.0 sPDF: VLX Seek 1.5 10B on Arc B580 report
RTX 40706-bit80K45 tok/s0.6 sPDF: VLX Seek 1.5 10B on RTX 4070 report
RTX 4070 Super6-bit80K45 tok/s0.5 sPDF: VLX Seek 1.5 10B on RTX 4070 Super report
RTX 50706-bit80K60 tok/s0.6 sPDF: VLX Seek 1.5 10B on RTX 5070 report
RTX 4060 Ti 16 GB8-bit128K21 tok/s0.8 sPDF: VLX Seek 1.5 10B on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit128K32 tok/s0.8 sPDF: VLX Seek 1.5 10B on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit128K28 tok/s0.9 sPDF: VLX Seek 1.5 10B on Arc A770 16 GB report
Radeon RX 9070 XT8-bit128K38 tok/s0.5 sPDF: VLX Seek 1.5 10B on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit128K47 tok/s0.4 sPDF: VLX Seek 1.5 10B on RTX 4070 Ti Super report
RTX 4080 Super8-bit128K51 tok/s0.3 sPDF: VLX Seek 1.5 10B on RTX 4080 Super report
RTX 5070 Ti8-bit128K61 tok/s0.4 sPDF: VLX Seek 1.5 10B on RTX 5070 Ti report
RTX 50808-bit128K66 tok/s0.3 sPDF: VLX Seek 1.5 10B on RTX 5080 report
Radeon RX 7900 XTX8-bit256K55 tok/s0.8 sPDF: VLX Seek 1.5 10B on Radeon RX 7900 XTX report
RTX 30908-bit256K64 tok/s0.5 sPDF: VLX Seek 1.5 10B on RTX 3090 report
RTX 40908-bit256K69 tok/s0.2 sPDF: VLX Seek 1.5 10B on RTX 4090 report
Radeon AI PRO R97008-bit256K38 tok/s0.5 sPDF: VLX Seek 1.5 10B on Radeon AI PRO R9700 report
RTX 50908-bit256K116 tok/s0.2 sPDF: VLX Seek 1.5 10B on RTX 5090 report
2× RTX 30908-bit256K64 tok/s0.3 sPDF: VLX Seek 1.5 10B on 2× RTX 3090 report
2× RTX 40908-bit256K69 tok/s0.1 sPDF: VLX Seek 1.5 10B on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit256K89 tok/s0.3 sPDF: VLX Seek 1.5 10B on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit256K66 tok/s0.2 sPDF: VLX Seek 1.5 10B on RTX 6000 Ada report
L40S8-bit256K59 tok/s0.3 sPDF: VLX Seek 1.5 10B on L40S report
2× RTX 50908-bit256K116 tok/s0.1 sPDF: VLX Seek 1.5 10B on 2× RTX 5090 report
4× RTX 30908-bit256K64 tok/s0.2 sPDF: VLX Seek 1.5 10B on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit256K116 tok/s0.1 sPDF: VLX Seek 1.5 10B on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB6-bit48K13 tok/s11 sPDF: VLX Seek 1.5 10B on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit136K10.0 tok/s11 sPDF: VLX Seek 1.5 10B on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit256K10.0 tok/s11 sPDF: VLX Seek 1.5 10B on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit136K22 tok/s5.7 sPDF: VLX Seek 1.5 10B on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit256K22 tok/s5.7 sPDF: VLX Seek 1.5 10B on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit256K22 tok/s5.7 sPDF: VLX Seek 1.5 10B on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit256K38 tok/s2.8 sPDF: VLX Seek 1.5 10B on M4 Max 128 GB report
M5 Max 128 GB8-bit256K43 tok/s0.8 sPDF: VLX Seek 1.5 10B on M5 Max 128 GB report
M3 Ultra 96 GB8-bit256K48 tok/s1.6 sPDF: VLX Seek 1.5 10B on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit256K48 tok/s1.6 sPDF: VLX Seek 1.5 10B on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit256K48 tok/s1.6 sPDF: VLX Seek 1.5 10B on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit256K67 tok/s0.5 sPDF: VLX Seek 1.5 10B on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit256K16 tok/s1.6 sPDF: VLX Seek 1.5 10B on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit256K16 tok/s1.6 sPDF: VLX Seek 1.5 10B on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit256K21 tok/s0.6 sPDF: VLX Seek 1.5 10B on RTX Spark report
DGX Spark8-bit256K20 tok/s0.6 sPDF: VLX Seek 1.5 10B on DGX Spark report
Dual DGX Spark8-bit256K20 tok/s0.3 sPDF: VLX Seek 1.5 10B on Dual DGX Spark report
DGX Station GB3008-bit256K345 tok/s0.1 sPDF: VLX Seek 1.5 10B on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit256K116 tok/s0.1 sPDF: VLX Seek 1.5 10B on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit256K116 tok/s0.1 sPDF: VLX Seek 1.5 10B on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit256K116 tok/s0.1 sPDF: VLX Seek 1.5 10B on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit256K130 tok/s0.4 sPDF: VLX Seek 1.5 10B on A100 80 GB SXM report
H100 SXM8-bit256K198 tok/s0.1 sPDF: VLX Seek 1.5 10B on H100 SXM report
4× H200 NVL8-bit256K262 tok/s0.1 sPDF: VLX Seek 1.5 10B on 4× H200 NVL report
8× H200 NVL8-bit256K262 tok/s0.1 sPDF: VLX Seek 1.5 10B 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 Asher-1/VLX-Seek-1.5-10B-GGUF
vLLM
vllm serve omlab/VLX-Seek-1.5-10B

Questions

How much VRAM does VLX Seek 1.5 10B need?

VLX Seek 1.5 10B needs 5.7 GB for its weights at 4-bit, 7.0 GB with an 8K context and 7.8 GB with 32K. At 8-bit, 11.3 GB.

What is the minimum VRAM for VLX Seek 1.5 10B?

About 5.9 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 VLX Seek 1.5 10B?

10B parameters: 18.7 GB of weights at 16-bit, 5.7 GB at 4-bit.

Can VLX Seek 1.5 10B run on an RTX 4090?

Yes. At 8-bit it holds 256K of context and writes about 69 tokens a second.

More models