Qwen3-0.6B

Alibaba · 0.8B · ctx 40K · apache-2.0 · HuggingFace

Qwen3-0.6B (Alibaba) is a 0.8B parameters open-weight model tracked on AI Hashrate. At 8K context we estimate roughly 1.6 GB VRAM for Q4 (weights 0.44 GB + KV 0.16 GB + 1 GB overhead) and 2.76 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 ≈ 5000 tok/s (estimated); MI300X 192GB ≈ 4215.9 tok/s (estimated); H200 141GB SXM5 ≈ 3818.2 tok/s (estimated); H100 80GB SXM5 ≈ 2664.8 tok/s (estimated); Gaudi 3 128GB ≈ 2605.9 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.