community pool / model answer

Laguna-XS-2.1-GGUF

chat · poolside/Laguna-XS-2.1-GGUF

103.9 tok/s · model median · 25 runs in the model record.

Reference rig	Quantization	Median result	Basis	TTFT (ms)	Peak VRAM	Max context	Engines
M1 Max 64GB	4-bit	62.3 tok/s	reported	705	- GB	2,048	llama.cpp
RTX 3090 24GB	3-bit	110.7 tok/s	measured	-	19.8 GB	262,144	llama.cpp
RTX 3090 24GB	4-bit	125.2 tok/s	measured	-	21.5 GB	131,072	llama.cpp
RTX 4070 12GB	4-bit	55.1 tok/s	reported	72,847	- GB	2,048	llama.cpp
RX 570 4GB ×2	4-bit	16.8 tok/s	reported	206	- GB	2,048	llama.cpp
Ryzen AI Max 395 64GB	4-bit	92.4 tok/s	reported	961	19.8 GB	8,192	llama.cpp

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.