MiniCPM5 2B
MiniCPM5 2B is an open-weights language model from Openbmb, released September 6, 2026. It has 2.5B parameters (dense). The context window is 128K tokens.
Check it against your own PC →
Can you run MiniCPM5 2B locally?
The smallest common GPU it runs on is 6 GB of VRAM (with 32 GB of RAM): Q6_K, fully on GPU, roughly 53 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~53 |
| 8 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~73 |
| 12 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~95 |
| 16 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~130 |
| 24 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~210 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.4 GB | – | ~340 |
Ready-made quantizations on Hugging Face: Abiray/MiniCPM5-2B-heretic-abliterated-GGUF · bartowski/MiniCPM5-2B-GGUF · peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF · lightonai/colgrep-agent-2B-GGUF · saidutta69/MiniCPM5-2B-heretic · SonGT1/MiniCPM5-abliterated-2B
Benchmark results
Epoch AI has not published results for MiniCPM5 2B yet. New models are usually evaluated within days to weeks.