AMD Instinct MI250X~433 tok/s (Q4)

Alibaba-NLP/gte-Qwen2-1.5B-instruct speed on AMD Instinct MI250X

Quantization-specific throughput and VRAM requirements for Alibaba-NLP/gte-Qwen2-1.5B-instruct running on AMD Instinct MI250X.

Speed Snapshot

Topline estimate from compatibility data

ModelAlibaba-NLP/gte-Qwen2-1.5B-instruct

GPUAMD Instinct MI250X

Q4 speed433 tok/s

Q4 VRAM required3GB

Data Source

Calculation and benchmark status

Speed values come from the compatibility dataset (`estimatedTokensPerSec`) and are sorted by quantization.

For full verdict logic and alternate GPUs, see the canonical compatibility page.

Quantization Speed Table

Quantization	VRAM needed	VRAM available	Speed	Verdict
Q4	3GB	128GB	433 tok/s	✅ Fits
Q8	5GB	128GB	329 tok/s	✅ Fits
FP16	11GB	128GB	191 tok/s	✅ Fits