MiniCPM O 2 6 — Hardware Requirements & GPU Compatibility
VisionFunctions
MiniCPM-o 2.6 is OpenBMB's 8.7-billion-parameter omni-modal model, combining a SigLip-400M vision encoder, a Whisper-medium audio encoder, a ChatTTS speech decoder, and a Qwen2.5-7B language backbone trained end to end. Unlike a pure vision-language model, it takes images, video, and audio as input and can generate spoken output too, supporting real-time bilingual speech conversation, voice cloning, and live-streaming understanding. At this size it is practical on a single high-end consumer GPU once quantized. Context length is 32,768 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in January 2025. Its distinguishing feature is the end-to-end omni-modal architecture, using time-division multiplexing that lets it process continuous video and audio streams and reply with synthesized speech, not just text.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 32K)Best on 24 GB (Q4_K_M · ~117 tok/s · 32K)Macs from 8 GB (IQ2_M · ~21 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run MiniCPM O 2 6
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/openbmb/MiniCPM-o-2_6-gguf:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 17.6 GB free · 16K context
Download: Model-7.6B-Q4_K_M.gguf4.68 GBopenbmb
Ways to runOther ways
each opens step-by-stepRun in cloud
Your own private copy · Q4_K_M · 24 GB GPU
- The easiest way to try this model
- Running in a few minutes, easy setup
- Chat right here or connect any app
$5+ bonus for new users
Runs on RTX 3080 Ti (12 GB) (5 GB headroom) · Q4_K_M
- Generation speed
- ~106 tok/s
- generation speed
- Cost per 1M output tokens
- $0.32
- per 1M output tokens
How Much VRAM Does MiniCPM O 2 6 Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 4.1 GB | 5.9 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 32K context · ~160 tok/s | |
| 4.2 GB | 6 GB | 3-bit small — 3-bit small quantization | Fits: 32K context · ~156 tok/s | |
| 4.7 GB | 6.4 GB | 3-bit medium — 3-bit medium quantization | Fits: 32K context · ~141 tok/s | |
| 4.8 GB | 6.5 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 32K context · ~138 tok/s | |
| 5.6 GB | 7.4 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 32K context · ~117 tok/s | |
| 6.6 GB | 8.4 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 32K context · ~99 tok/s | |
| 7.6 GB | 9.3 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 32K context · ~87 tok/s | |
| 9.1 GB | 10.9 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 32K context · ~72 tok/s |
Selected Q4_K_M
4-bit medium quantization6.4GBat 16K context
- Weights
- 5.2 GB
- KV cache
- 0.9 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 17.6 GB free · ~117 tok/s
≥50 tok/s needs ≥432 GB/s · RTX 5090 ~207, RTX 3090 Ti ~117
Which Hardware Can Run MiniCPM O 2 6?
Q4_K_M · 5.6 GBMiniCPM O 2 6 (Q4_K_M) requires 5.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 8+ GB is recommended. Using the full 33K context window can add up to 1.8 GB, bringing total usage to 7.4 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti. 151 devices with unified memory can run MiniCPM O 2 6, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).
≤ 11 GB
≤ 11 GB tier →fits · 1.6 GB free · up to 32K ctx
16 GPUs · 1 device family
RTX 3080~88Simulate output on RTX 3080RTX 2080 Ti~71Simulate output on RTX 2080 TiRTX 3070 Ti~70Simulate output on RTX 3070 TiOrin Nano~12Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)20 GB
fits · 13.6 GB free · up to 32K ctx
1 GPU
RX 7900 XT~85Simulate output on RX 7900 XT32–40 GB
fits · 25.6 GB free · up to 32K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~207Simulate output on RTX 5090M5 Max36 GB~57Simulate output on Mac Studio M5 Max (36 GB)M532 GB~19Simulate output on MacBook Air 13" M5 (32 GB)Orin~24Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~16Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to MiniCPM O 2 6
MiniCPM family · 18 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Granite Switch 4.1 8BIBM | — | 8.4 GB | ~111 | same size |
| ALoDLM 8BAmazon | — | 7.6 GB | ~119 | same size |
| Llm Jp 4.1 8B Thinkingllm-jp | — | 7.6 GB | ~115 | same size |
| Apertus V1.5 8Bswiss-ai | — | 8.5 GB | ~111 | same size |
| LFM2.5 8B A1BLiquid AI | — | 5.6 GB | ~335 | same size |
| Chandradatalab-to | — | 8 GB | ~112 | same size |
Specifications
- Publisher
- OpenBMB
- Family
- MiniCPM
- Parameters
- 8.7B
- Architecture
- MiniCPMO
- Context Length
- 32,768 tokens
- Vocabulary Size
- 151,700
- Release Date
- 2025-01-12
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does MiniCPM O 2 6 need?
MiniCPM O 2 6 requires 5.6 GB of VRAM at Q4_K_M, or 17.8 GB at BF16. Full 33K context adds up to 1.8 GB (7.4 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 8.7B × 4.8 bits ÷ 8 = 5.2 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 2.2 GB (at full 33K context)
VRAM usage by quantization
What's the best quantization for MiniCPM O 2 6?
For MiniCPM O 2 6, Q4_K_M (5.6 GB) offers the best balance of quality and VRAM usage. Q4_K_L (5.7 GB) provides better quality if you have the VRAM. The smallest option is IQ2_M at 3.4 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run MiniCPM O 2 6 on a Mac?
Yes — Mac Mini M2 (8 GB) and 145 other Macs can run MiniCPM O 2 6, the smallest at IQ2_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 5.6 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M2 Pro (16 GB).
Can I run MiniCPM O 2 6 locally?
Yes — MiniCPM O 2 6 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is MiniCPM O 2 6?
At Q4_K_M, MiniCPM O 2 6 can reach ~854 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~117 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 ÷ 5.62 × 0.65 = ~925 tok/s
Estimated speed at Q4_K_M (5.6 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of MiniCPM O 2 6?
At Q4_K_M, the download is about 5.20 GB. The full-precision BF16 version is 17.35 GB. The smallest option (IQ2_M) is 2.93 GB.
Which GPUs can run MiniCPM O 2 6?
52 consumer GPUs can run MiniCPM O 2 6 at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.
Which devices can run MiniCPM O 2 6?
151 devices with unified memory can run MiniCPM O 2 6 at Q4_K_M (5.6 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (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.