Skip to content

Teutonic I Reliquary RL VRAM requirements

How much VRAM Teutonic I Reliquary RL needs
4-bit weights: 5.1 GBWith 8K context: 6.4 GBWith 32K context: 7.2 GB8-bit, 8K context: 10.2 GB
Parameters
9B
Longest context
256K
Released
Oct 2026
License
Apache 2.0
Strongest at
Not rated

How much VRAM Teutonic I Reliquary RL needs

Memory Teutonic I Reliquary RL needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
16-bit16.7 GB18.0 GB18.8 GB21.8 GB
8-bit8.9 GB10.2 GB11.0 GB14.0 GB
6-bit6.8 GB8.2 GB8.9 GB11.9 GB
5-bit5.9 GB7.3 GB8.0 GB11.0 GB
4-bit5.1 GB6.4 GB7.2 GB10.2 GB
3-bit4.1 GB5.4 GB6.2 GB9.2 GB
2-bit3.1 GB4.5 GB5.2 GB8.2 GB
1.6-bit1.7 GB3.1 GB3.8 GB6.8 GB

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

GPUs and Macs that run Teutonic I Reliquary RL

The best way Teutonic I Reliquary RL runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 40605-bit16K32 tok/s1.0 sPDF: Teutonic I Reliquary RL on RTX 4060 report
RTX 50605-bit16K52 tok/s0.8 sPDF: Teutonic I Reliquary RL on RTX 5060 report
RTX 30705-bit16K52 tok/s0.8 sPDF: Teutonic I Reliquary RL on RTX 3070 report
RTX 3080 10 GB6-bit48K75 tok/s0.5 sPDF: Teutonic I Reliquary RL on RTX 3080 10 GB report
RTX 3060 12 GB8-bit44K29 tok/s1.2 sPDF: Teutonic I Reliquary RL on RTX 3060 12 GB report
Arc B5808-bit44K26 tok/s0.8 sPDF: Teutonic I Reliquary RL on Arc B580 report
RTX 40708-bit44K40 tok/s0.5 sPDF: Teutonic I Reliquary RL on RTX 4070 report
RTX 4070 Super8-bit44K40 tok/s0.4 sPDF: Teutonic I Reliquary RL on RTX 4070 Super report
RTX 50708-bit44K53 tok/s0.5 sPDF: Teutonic I Reliquary RL on RTX 5070 report
RTX 4060 Ti 16 GB8-bit160K23 tok/s0.7 sPDF: Teutonic I Reliquary RL on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB8-bit160K36 tok/s0.7 sPDF: Teutonic I Reliquary RL on RTX 5060 Ti 16 GB report
Arc A770 16 GB8-bit160K31 tok/s0.7 sPDF: Teutonic I Reliquary RL on Arc A770 16 GB report
Radeon RX 9070 XT8-bit160K42 tok/s0.4 sPDF: Teutonic I Reliquary RL on Radeon RX 9070 XT report
RTX 4070 Ti Super8-bit160K53 tok/s0.4 sPDF: Teutonic I Reliquary RL on RTX 4070 Ti Super report
RTX 4080 Super8-bit160K57 tok/s0.3 sPDF: Teutonic I Reliquary RL on RTX 4080 Super report
RTX 5070 Ti8-bit160K69 tok/s0.4 sPDF: Teutonic I Reliquary RL on RTX 5070 Ti report
RTX 50808-bit160K74 tok/s0.3 sPDF: Teutonic I Reliquary RL on RTX 5080 report
Radeon RX 7900 XTX8-bit256K61 tok/s0.7 sPDF: Teutonic I Reliquary RL on Radeon RX 7900 XTX report
RTX 30908-bit256K72 tok/s0.4 sPDF: Teutonic I Reliquary RL on RTX 3090 report
RTX 40908-bit256K77 tok/s0.2 sPDF: Teutonic I Reliquary RL on RTX 4090 report
Radeon AI PRO R97008-bit256K42 tok/s0.4 sPDF: Teutonic I Reliquary RL on Radeon AI PRO R9700 report
RTX 50908-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on RTX 5090 report
2× RTX 30908-bit256K72 tok/s0.2 sPDF: Teutonic I Reliquary RL on 2× RTX 3090 report
2× RTX 40908-bit256K77 tok/s0.1 sPDF: Teutonic I Reliquary RL on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit256K100 tok/s0.2 sPDF: Teutonic I Reliquary RL on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit256K74 tok/s0.2 sPDF: Teutonic I Reliquary RL on RTX 6000 Ada report
L40S8-bit256K67 tok/s0.3 sPDF: Teutonic I Reliquary RL on L40S report
2× RTX 50908-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on 2× RTX 5090 report
4× RTX 30908-bit256K72 tok/s0.1 sPDF: Teutonic I Reliquary RL on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on RTX PRO 6000 Blackwell report
Mac mini M4 16 GB6-bit72K14 tok/s9.8 sPDF: Teutonic I Reliquary RL on Mac mini M4 16 GB report
Mac mini M4 24 GB8-bit168K11 tok/s9.8 sPDF: Teutonic I Reliquary RL on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit256K11 tok/s9.8 sPDF: Teutonic I Reliquary RL on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB8-bit168K25 tok/s4.9 sPDF: Teutonic I Reliquary RL on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB8-bit256K25 tok/s4.9 sPDF: Teutonic I Reliquary RL on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit256K25 tok/s4.9 sPDF: Teutonic I Reliquary RL on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit256K43 tok/s2.5 sPDF: Teutonic I Reliquary RL on M4 Max 128 GB report
M5 Max 128 GB8-bit256K48 tok/s0.7 sPDF: Teutonic I Reliquary RL on M5 Max 128 GB report
M3 Ultra 96 GB8-bit256K54 tok/s1.4 sPDF: Teutonic I Reliquary RL on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit256K54 tok/s1.4 sPDF: Teutonic I Reliquary RL on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit256K54 tok/s1.4 sPDF: Teutonic I Reliquary RL on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit256K74 tok/s0.4 sPDF: Teutonic I Reliquary RL on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit256K18 tok/s1.4 sPDF: Teutonic I Reliquary RL on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit256K18 tok/s1.4 sPDF: Teutonic I Reliquary RL on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit256K24 tok/s0.5 sPDF: Teutonic I Reliquary RL on RTX Spark report
DGX Spark8-bit256K22 tok/s0.5 sPDF: Teutonic I Reliquary RL on DGX Spark report
Dual DGX Spark8-bit256K22 tok/s0.3 sPDF: Teutonic I Reliquary RL on Dual DGX Spark report
DGX Station GB3008-bit256K375 tok/s0.1 sPDF: Teutonic I Reliquary RL on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit256K129 tok/s0.1 sPDF: Teutonic I Reliquary RL on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit256K145 tok/s0.3 sPDF: Teutonic I Reliquary RL on A100 80 GB SXM report
H100 SXM8-bit256K218 tok/s0.1 sPDF: Teutonic I Reliquary RL on H100 SXM report
4× H200 NVL8-bit256K287 tok/s0.1 sPDF: Teutonic I Reliquary RL on 4× H200 NVL report
8× H200 NVL8-bit256K287 tok/s0.1 sPDF: Teutonic I Reliquary RL 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/ReliquaryForge_Teutonic-I-Reliquary-RL-GGUF
vLLM
vllm serve ReliquaryForge/Teutonic-I-Reliquary-RL

Questions

How much VRAM does Teutonic I Reliquary RL need?

Teutonic I Reliquary RL needs 5.1 GB for its weights at 4-bit, 6.4 GB with an 8K context and 7.2 GB with 32K. At 8-bit, 10.2 GB.

What is the minimum VRAM for Teutonic I Reliquary RL?

About 5.4 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 Teutonic I Reliquary RL?

9B parameters: 16.7 GB of weights at 16-bit, 5.1 GB at 4-bit.

Can Teutonic I Reliquary RL run on an RTX 4090?

Yes. At 8-bit it holds 256K of context and writes about 77 tokens a second.

More models