Home / Models / MiniCPM5-2B

How to Run MiniCPM5-2B Locally — Setup Guide (September 2026)

MiniCPM5-2B is a dense 2.5B model built for on-device deployment, and the second release in OpenBMB's MiniCPM5 line. Independent benchmarking (Artificial Analysis) puts it ahead of every other open-weight model under 4B parameters on their Intelligence Index — a real result for a model this small, not a vendor-only claim.

Numbers you can trust

We don't print guessed speeds. Every measured number on this page came from TurboLLM's own auto-tuner running MiniCPM5-2B on real hardware — labeled as such. When you run it yourself, TurboLLM benchmarks it on your exact GPU and shows a VRAM-fit verdict before you load, and the real measured tokens/sec once it's running.

Does it fit your GPU?

2.5B dense parameters, conventional full attention on all 42 layers (GQA, 16 query heads / 2 KV heads), 131K native context. Ships with a hybrid reasoning toggle — the same checkpoint switches between a fast no-think mode and a deliberate think mode via the chat template's enable_thinking flag, rather than needing two separate downloads.

MiniCPM5-2B
OpenBMB · GGUF
Dense · full attention
toolsthinkingcode
VRAMFit
4 GB / CPU onlyQ4_K_M (1.56 GB) — runs on integrated graphics or plain CPU at usable speed · 131K context
8 GBQ8_0 (2.68 GB) fully in VRAM, near-lossless, with plenty of room for context or an embedding model alongside it · 131K context
openbmb/MiniCPM5-2B-GGUF
View on Hugging Face →

How to run it

  1. Install TurboLLM

    npx turbollm — no install step, works on Windows, macOS, and Linux. It detects your GPU and auto-provisions a matching llama-server build (no CUDA toolkit, no Python, no compiler).

  2. Get the model

    Paste openbmb/MiniCPM5-2B-GGUF into TurboLLM's in-app Hugging Face search — it lists every quant with a VRAM-fit verdict against your real free VRAM, and downloads are resumable and SHA-256 verified.

  3. Let auto-fit pick the setup

    Flip the auto-fit toggle and TurboLLM chooses the GPU/CPU layer split (and, for a Mixture-of-Experts model, the expert-offload split) for your card — the same manual flag-hunting every other guide walks you through by hand.

  4. Load it

    TurboLLM benchmarks the model on your exact GPU as it loads and shows the real measured tokens/sec — not a number copied from someone else's card.

  5. Use it

    Chat in the built-in UI, or point any OpenAI- or Anthropic-compatible tool (including Claude Code) at TurboLLM's local API — see the API docs.

One honest caveat

The thinking-mode toggle lives in the chat template's enable_thinking flag. OpenBMB's own GGUF ships the template with it wired up, but a re-quant from a different uploader may not carry it — if thinking mode won't switch off, check that the GGUF you loaded came from the first-party openbmb/MiniCPM5-2B-GGUF repo.

FAQ

Which quant should I use for MiniCPM5-2B?

Pick the highest quant that fits your VRAM with room left over for the KV cache and compute buffer — not just the model weights. As a starting point, Q4_K_M is usually the sweet spot (small quality loss for a large size drop); step up to Q5_K_M/Q6_K/Q8_0 if you have headroom, or down to Q3_K_M/an IQ quant if you don't. TurboLLM shows a VRAM-fit verdict for every quant against your card's real free VRAM before you download — see Quantization explained for what the letters and numbers actually mean, or llama.cpp flags that actually matter for the other settings worth knowing.

Will MiniCPM5-2B run on my GPU?

Check the VRAM table above for the tier closest to your card. If you're not sure which tier you're in, or want picks tailored to your exact hardware, use What can I run? — enter your VRAM (or Apple unified memory) and it suggests the model, quant, and offload split that fits, the same logic TurboLLM's auto-fit runs when you load a model.

How is running MiniCPM5-2B on TurboLLM different from Ollama or LM Studio?

All three can run MiniCPM5-2B as a GGUF. The differences show up around the edges: TurboLLM auto-tunes the quant and GPU/CPU split against your card's measured free VRAM rather than a fixed default, shows real generation speed instead of a guess, and lets you add a community fork — like the one behind this page, where relevant — without compiling anything yourself. Ollama and LM Studio each ship one blessed runtime you can't swap out. If your current setup already works well for you, there's no need to switch.

Run it now

$ npx turbollm

One command detects your GPU, provisions the right engine, and opens the UI. New here? Start with Install & first run and Quantization explained. On different hardware, see RTX 4060 or Apple M4 (16 GB). Also see how to run it or how to run it.