Qwen2.5-7B-Instruct

Alibaba · 7.6B · ctx 128K · apache-2.0 · HuggingFace

Qwen2.5-7B-Instruct (Alibaba) is a 7.6B parameters open-weight model tracked on AI Hashrate. At 8K context we estimate roughly 6.7 GB VRAM for Q4 (weights 4.18 GB + KV 1.52 GB + 1 GB overhead) and 17.72 GB for FP16. About 50 GPUs/accelerators in our catalog fully fit this model at Q4 under that context assumption. Fastest Q4 configs: B200 192GB SXM ≈ 669.9 tok/s (estimated); MI300X 192GB ≈ 443.8 tok/s (estimated); H200 141GB SXM5 ≈ 401.9 tok/s (estimated); H100 80GB SXM5 ≈ 280.5 tok/s (estimated); Gaudi 3 128GB ≈ 274.3 tok/s (estimated). Use the table below for VRAM fit and tok/s per dollar (list MSRP). Estimates follow memory-bandwidth math; measured rows override when present. See methodology for details. Methodology.

100 hardware configs — sorted by tok/s. Default fit context 8K.