Mistral-7B-Instruct-v0.3

Mistral · 7.2B · ctx 32K · Apache 2.0 · HuggingFace

Mistral-7B-Instruct-v0.3 (Mistral) is a 7.2B parameters open-weight model tracked on AI Hashrate. At 8K context we estimate roughly 6.4 GB VRAM for Q4 (weights 3.96 GB + KV 1.44 GB + 1 GB overhead) and 16.84 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 ≈ 707.1 tok/s (estimated); MI300X 192GB ≈ 468.4 tok/s (estimated); H200 141GB SXM5 ≈ 424.2 tok/s (estimated); H100 80GB SXM5 ≈ 296.1 tok/s (estimated); Gaudi 3 128GB ≈ 289.5 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.