MiniCPM5 2B VRAM requirements

How much VRAM does MiniCPM5 2B need?

MiniCPM5 2B needs 2.42 GB at Q4_K_M, 3.60 GB at Q8_0 and 6.02 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.

Fine-tuning MiniCPM5 2B? MiniCPM5 2B VRAM for LoRA, QLoRA and full training.

MiniCPM5 2B GGUF files on Hugging Face

The 8 GGUF files llama.cpp picks for MiniCPM5 2B with -hf <repo>:<QUANT>, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.

GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 1.6 GB, in 10⁹ bytes.

Which MiniCPM5 2B GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac

From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (1.50 GB). One request, FP16 KV cache.

Run MiniCPM5 2B with llama.cpp

llama-server -hf bartowski/MiniCPM5-2B-GGUF:Q4_K_M -c 131072

MiniCPM5-2B-Q4_K_M.gguf, 1.6 GB, from bartowski/MiniCPM5-2B-GGUF (checked 2026-09-29). On a 24 GB RTX 3090 / 4090, Q4_K_M with -c 131072 takes 7.93 GB and leaves 16.1 GB.

MiniCPM5 2B at 8K, 32K and 128K (full) tokens of context

All 42 layers use full attention. Each extra token of context adds 42 KB of FP16 cache per request. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.

How fast MiniCPM5 2B writes

Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth. A dash means it does not fit on one card. Try other settings in the speed calculator.

Longest context on one GPU

How many tokens of context MiniCPM5 2B fits on a single card with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.

Model details

Parameters

2.5B (2,516,756,480)

Layers

All 42 layers use full attention

Attention cache

2 KV heads × 128

Context length

131,072 tokens

Published weights

4.69 GB (BF16)

On Hugging Face

openbmb/MiniCPM5-2B

Reported benchmarks

MMLU-Pro 70.8 · LiveCodeBench v6 69.1. Vendor-reported: from the maker’s own model card, not rerun here; read 2026-09-30.

MiniCPM5 2B next to similar-sized models

The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.