community pool / model answer

Qwen3.8-Flash-Next-GGUF

chat · unsloth/Qwen3.8-Flash-Next-GGUF

44 tok/s · model median · 17 runs in the model record.

Reference rig	Quantization	Median result	Basis	TTFT (ms)	Peak VRAM	Max context	Engines
CMP 170HX 64GB	1-bit	44.6 tok/s	measured	321	42.1 GB	65,536	llama.cpp
GB10 Grace Blackwell 128GB	4-bit	27.3 tok/s	reported	-	- GB	131,072	llama.cpp
M5 Max 128GB	3-bit	41.5 tok/s	reported	289	- GB	2,048	llama.cpp
M5 Max 128GB	4-bit	39.4 tok/s	reported	304	- GB	2,048	llama.cpp
Radeon AI Pro R9700 32GB	4-bit	25.6 tok/s	reported	4,747	32.9 GB	32,768	llama.cpp
RTX 3090 24GB ×2	3-bit	41.6 tok/s	reported	6,505	43.6 GB	217,088	llama.cpp
RTX 5090 32GB ×2	4-bit	142 tok/s	reported	515	59.9 GB	131,072	llama.cpp
Ryzen AI Max 395 128GB	4-bit	74.9 tok/s	measured	1,356	68.3 GB	32,768	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.