Can I run MiniCPM5-2B locally?

Yes—its official 1.6GB Q4_K_M GGUF fits on modern phones, Macs, and PCs. At Q4_K_M, the estimated 4K working set is ~2.5GB. It is a dense 2.52B bilingual model aimed at local assistants, coding, and tool use. Basic chat works through llama.cpp or Ollama; reliable agent tool execution still needs a compatible tool-calling harness.

Params
2.52B
Family
minicpm
Released
2026-09
Tags
chat · coding · reasoning

How to run MiniCPM5-2B locally

OpenBMB publishes this GGUF and the commands below. Start at 4K context, confirm a clean text response, then increase context while watching memory. The advertised 128K maximum is a model capability—not a promise that a low-memory phone can allocate the full KV cache. For native tool calling, OpenBMB recommends SGLang with its minicpm5 tool parser.

MAC OR PC · LLAMA.CPP
llama serve -hf openbmb/MiniCPM5-2B-GGUF:Q4_K_M
# Or with Ollama:
ollama run hf.co/openbmb/MiniCPM5-2B-GGUF:Q4_K_M
OpenBMB official GGUF files and instructions

Memory figures on this page are planning estimates from GGUF bytes, 5% weight overhead, a 4K KV cache, and runtime allowance—not measured peak allocation or a speed guarantee.

Downloads & memory by quant

The download is only the weights — add KV cache and runtime to get what your phone actually needs. “Memory-fit” only means it can load; “usable” also requires an estimated 3+ tokens/s. Each row uses that row's quant. ★ = recommended.

QuantDownloadMemory neededPhone reality
Q4_K_M1.6 GB~2.5 GB139 memory-fit · 139 usable
Q8_02.7 GB~3.6 GB139 memory-fit · 127 usable

MiniCPM5-2B at Q4_K_M on 140 phones

This table is fixed to the recommended Q4_K_M, on each phone's largest RAM option. Counts for smaller quants above will differ. All phones are shown, sorted by fit and estimated experience; tap one for the full report.

PhoneChip · RAMNeedsSpeedVerdict

~ = bandwidth-based estimate · ✓ = measured on real hardware

MiniCPM5-2B on 54 Mac configurations

Each Mac uses the recommended Q4_K_M at 4K context when it fits, or the largest smaller tracked quant that fits. Each row links to the full report with a macOS-specific setup guide.

MacUnified memoryQuantWorking budgetNeedsEstimated speedVerdict

FAQ

How much RAM do I need to run MiniCPM5-2B on a phone?

At Q4_K_M, MiniCPM5-2B needs ~2.5GB of usable memory (weights + KV cache + runtime). In practice that means a 6GB+ Android phone or a 6GB+ iPhone — iOS lets a single app use less of its RAM than Android does.

How fast does MiniCPM5-2B run on a flagship phone?

On the Galaxy S25 Ultra (Snapdragon 8 Elite) it runs at ~21.6 tokens/s at Q4_K_M, estimated from memory bandwidth. Anything above ~8 tokens/s feels smooth for chat.

Can an iPhone run MiniCPM5-2B?

Yes — the iPhone 16 Pro Max runs it at ~16.9 tokens/s at Q4_K_M, using 2.5GB of its ~5.2GB usable memory.

What is the best quantization of MiniCPM5-2B for mobile?

Q4_K_M (1.6GB download) is the size/quality sweet spot of the 2 quants available. Total memory needed is ~2.5GB once the KV cache and runtime are counted — the download size alone understates it.

How many phones can run MiniCPM5-2B?

139 of the 140 phones we track have enough estimated usable memory at Q4_K_M; 139 also reach our usable-speed threshold of 3 tokens/s, and 106 run smoothly. Memory-fit alone does not mean a good experience.

Which Macs can run MiniCPM5-2B?

54 of the 54 exact Mac configurations we track have enough estimated working memory. Each Mac uses the recommended Q4_K_M when it fits, or the largest smaller tracked quant that fits, and separates installed unified memory, the macOS reserve, and estimated speed.

What is MiniCPM5-2B good for on a phone?

It's tagged for chat, coding, reasoning. At 2.52B parameters, it's a fast, lightweight pick for quick tasks.

Similar models

LFM2.5 2.6BMinistral 3 3BG9v3 3BGranite 4.2 3BQwen 3.5 2B