MiniCPM5 2B — Hardware Requirements & GPU Compatibility
Chat
MiniCPM5-2B is the second model in OpenBMB's MiniCPM5 series, a dense 2.5-billion-parameter causal language model built on a Llama-style architecture. This release is the final version, post-trained with reinforcement learning and on-policy distillation, and is aimed at local assistants, coding agents, tool-use workflows, and reasoning tasks where a small footprint matters, including on-device and edge deployment. It supports a 131,072-token context window, unusually long for its size, and ships under the Apache 2.0 license, allowing free commercial and non-commercial use. At roughly 2.5 billion parameters, it needs well under 2GB of memory at 4-bit quantization, so it runs comfortably on almost any modern GPU or laptop.
564.4K downloads
1.7K likesSpecifications
- Publisher
- OpenBMB
- Family
- MiniCPM
- Parameters
- 2.5B
- Architecture
- LlamaForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 130,560
- Release Date
- 2026-09-06
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does MiniCPM5 2B Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 1.5 GB | 7.0 GB | 1.07 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 1.5 GB | 7.0 GB | 1.10 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 1.6 GB | 7.2 GB | 1.23 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 1.6 GB | 7.2 GB | 1.26 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 1.9 GB | 7.5 GB | 1.51 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 2.2 GB | 7.7 GB | 1.79 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 2.5 GB | 8.0 GB | 2.08 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 2.9 GB | 8.4 GB | 2.52 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run MiniCPM5 2B?
Q4_K_M · 1.9 GBMiniCPM5 2B (Q4_K_M) requires 1.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 3+ GB is recommended. Using the full 131K context window can add up to 5.6 GB, bringing total usage to 7.5 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run MiniCPM5 2B?
Q4_K_M · 1.9 GB59 devices with unified memory can run MiniCPM5 2B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomWhere to Download MiniCPM5 2B
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does MiniCPM5 2B need?
MiniCPM5 2B requires 1.9 GB of VRAM at Q4_K_M, or 5.4 GB at BF16. Full 131K context adds up to 5.6 GB (7.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 2.5B × 4.8 bits ÷ 8 = 1.5 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 6 GB (at full 131K context)
VRAM usage by quantization
Q4_K_M1.9 GBQ4_K_M + full context7.5 GB- What's the best quantization for MiniCPM5 2B?
For MiniCPM5 2B, Q4_K_M (1.9 GB) offers the best balance of quality and VRAM usage. Q4_K_L (1.9 GB) provides better quality if you have the VRAM. The smallest option is IQ2_M at 1.2 GB.
VRAM requirement by quantization
IQ2_M1.2 GBIQ3_M1.5 GBQ4_11.8 GBQ4_K_M ★1.9 GBQ4_K_L1.9 GBBF165.4 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run MiniCPM5 2B on a Mac?
MiniCPM5 2B requires at least 1.2 GB at IQ2_M, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run MiniCPM5 2B locally?
Yes — MiniCPM5 2B can run locally on consumer hardware. At Q4_K_M quantization it needs 1.9 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is MiniCPM5 2B?
At Q4_K_M, MiniCPM5 2B can reach ~2526 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~345 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 1.9 × 0.65 = ~2737 tok/s
Estimated speed at Q4_K_M (1.9 GB)
~2737 tok/s~345 tok/s~2737 tok/s~2526 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of MiniCPM5 2B?
At Q4_K_M, the download is about 1.51 GB. The full-precision BF16 version is 5.03 GB. The smallest option (IQ2_M) is 0.85 GB.
- Which GPUs can run MiniCPM5 2B?
52 consumer GPUs can run MiniCPM5 2B at Q4_K_M (1.9 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
- Which devices can run MiniCPM5 2B?
59 devices with unified memory can run MiniCPM5 2B at Q4_K_M (1.9 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.