GLM-4.7-Flash

GLM family

2recorded runs

182.5best tok/s

0verified

Evidence: 0 sourced runs · 2 awaiting review · 1 independent source · first recorded Aug 17, 2026 · last recorded Aug 19, 2026. What these labels mean.

What the GLM-4.7-Flash numbers mean

The fastest result on our board is 182.5 output tok/s on an RTX 5090 at Q4_K_M with a 2048-token context. That is about 5.5 ms per generated token — 500 tokens takes roughly 2.7 seconds. Prompt processing on the same record runs at 699.6 tok/s, so a 2,000-token prompt is read in about 2.9 seconds. Across the 2 records we hold for this model the range is 82.0–182.5 tok/s on 2 GPUs, because quantisation, context length and the engine build all move the number.

Fastest GPUs for this model (best tok/s)

RTX 5090182.5
Radeon AI Pro R970082.0

All benchmarks Methodology

GPU	Model	Quant	Backend	Ctx	tok/s	prompt tok/s	VRAM GB	Source
RTX 5090 · 32 GB	GLM-4.7-Flash	Q4_K_M	lmstudio —	2048	182.5 …	699.6	—	src
Radeon AI Pro R9700 · 32 GB	GLM-4.7-Flash	Q2_K_XL	llama.cpp —	2048	82.0 …	2046.4	—	src