gemma-3-4b-it

Google · 4.3B · ctx 128K · Gemma · HuggingFace

gemma-3-4b-it (Google) is a 4.3B parameters open-weight model tracked on AI Hashrate. At 8K context we estimate roughly 4.15 GB VRAM for Q4 (weights 2.37 GB + KV 0.78 GB + 1 GB overhead) and 10.38 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 ≈ 1305.4 tok/s (estimated); MI300X 192GB ≈ 864.8 tok/s (estimated); H200 141GB SXM5 ≈ 783.2 tok/s (estimated); H100 80GB SXM5 ≈ 546.6 tok/s (estimated); Gaudi 3 128GB ≈ 534.5 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.