# MiniMax-M2.5: How to Run Guide MiniMax-M2.5 is a new open LLM achieving SOTA in coding, agentic tool use and search and office work, scoring 80.2% in [SWE-Bench](#benchmarks) Verified, 51.3% in Multi-SWE-Bench, and 76.3% in BrowseComp. The **230B parameters** (10B active) model has a **200K context** window and unquantized bf16 requires **457GB**. Unsloth Dynamic **3-bit** GGUF reduces size to **101GB** **(-62%):** [**MiniMax-M2.5 GGUF**](https://huggingface.co/unsloth/MiniMax-M2.5-GGUF) All uploads use Unsloth [Dynamic 2.0](https://unsloth.ai/docs/basics/unsloth-dynamic-2.0-ggufs) for SOTA quantization performance - so 3-bit has important layers upcasted to 8 or 16-bit. You can also fine-tune the model via Unsloth, using multiGPUs. {% hint style="success" %} **Feb 26:** See how well our GGUF quants [perform on benchmarks here](#unsloth-gguf-benchmarks). {% endhint %} ### :gear: Usage Guide The 3-bit dynamic quant UD-Q3\_K\_XL uses **101GB** of disk space - this fits nicely on a **128GB unified memory Mac** for \~20+ tokens/s, and also works faster with a **1x16GB GPU and 96GB of RAM** for 25+ tokens/s. **2-bit** quants or the biggest 2-bit will fit on a 96GB device. For near **full precision**, use `Q8_0` (8-bit) which utilizes 243GB and will fit on a 256GB RAM device / Mac for 10+ tokens/s. {% hint style="success" %} For best performance, make sure your total available memory (VRAM + system RAM) exceeds the size of the quantized model file you’re downloading. If it doesn’t, llama.cpp can still run via SSD/HDD offloading, but inference will be slower. {% endhint %} ### Recommended Settings MiniMax recommends using the following parameters for best performance: `temperature=1.0`, `top_p = 0.95`, `top_k = 40`. {% columns %} {% column %} | Default Settings (Most Tasks) | | ---------------------------------- | | `temperature = 1.0` | | `top_p = 0.95` | | `top_k = 40` | | `repeat penalty = 1.0` or disabled | | {% endcolumn %} | {% column %} * **Maximum context window:** `196,608` * `Min_P = 0.01` (default might be 0.05) * Default system prompt: {% code overflow="wrap" %} ``` You are a helpful assistant. Your name is MiniMax-M2.5 and is built by MiniMax. ``` {% endcode %} {% endcolumn %} {% endcolumns %} ## Run MiniMax-M2.5 Tutorials: For these tutorials, we will be utilizing the 3-bit [UD-Q3\_K\_XL](https://huggingface.co/unsloth/MiniMax-M2.5-GGUF?show_file_info=UD-Q3_K_XL%2FMiniMax-M2.5-UD-Q3_K_XL-00001-of-00004.gguf) quant which fits in a 128GB RAM device. #### ✨ Run in llama.cpp {% stepper %} {% step %} Obtain the latest `llama.cpp` on [GitHub here](https://github.com/ggml-org/llama.cpp). You can follow the build instructions below as well. Change `-DGGML_CUDA=ON` to `-DGGML_CUDA=OFF` if you don't have a GPU or just want CPU inference. **For Apple Mac / Metal devices**, set `-DGGML_CUDA=OFF` then continue as usual - Metal support is on by default. {% code overflow="wrap" %} ```bash apt-get update apt-get install pciutils build-essential cmake curl libcurl4-openssl-dev -y git clone https://github.com/ggml-org/llama.cpp cmake llama.cpp -B llama.cpp/build \ -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON cmake --build llama.cpp/build --config Release -j --clean-first --target llama-cli llama-mtmd-cli llama-server llama-gguf-split cp llama.cpp/build/bin/llama-* llama.cpp ``` {% endcode %} {% endstep %} {% step %} If you want to use `llama.cpp` directly to load models, you can do the below: (:Q3\_K\_XL) is the quantization type. You can also download via Hugging Face (point 3). This is similar to `ollama run` . Use `export LLAMA_CACHE="folder"` to force `llama.cpp` to save to a specific location. Remember the model has only a maximum of 200K context length. Follow this for **most default** use-cases: ```bash export LLAMA_CACHE="unsloth/MiniMax-M2.5-GGUF" ./llama.cpp/llama-cli \ -hf unsloth/MiniMax-M2.5-GGUF:UD-Q3_K_XL \ --ctx-size 16384 \ --flash-attn on \ --temp 1.0 \ --top-p 0.95 \ --min-p 0.01 \ --top-k 40 ``` {% endstep %} {% step %} Download the model via (after installing `pip install huggingface_hub hf_transfer` ). You can choose `UD-Q3_K_XL` (dynamic 4-bit quant) or other quantized versions like `UD-Q6_K_XL` . We recommend using our 4bit dynamic quant `UD-Q3_K_XL` to balance size and accuracy. If downloads get stuck, see [hugging-face-hub-xet-debugging](https://unsloth.ai/docs/basics/troubleshooting-and-faqs/hugging-face-hub-xet-debugging "mention") ```bash hf download unsloth/MiniMax-M2.5-GGUF \ --local-dir unsloth/MiniMax-M2.5-GGUF \ --include "*UD-Q3_K_XL*" # Use "*Q8_0*" for 8-bit ``` {% endstep %} {% step %} You can edit `--threads 32` for the number of CPU threads, `--ctx-size 16384` for context length, `--n-gpu-layers 2` for GPU offloading on how many layers. Try adjusting it if your GPU goes out of memory. Also remove it if you have CPU only inference. {% code overflow="wrap" %} ```bash ./llama.cpp/llama-cli \ --model unsloth/MiniMax-M2.5-GGUF/UD-Q3_K_XL/MiniMax-M2.5-UD-Q3_K_XL-00001-of-00004.gguf \ --temp 1.0 \ --top-p 0.95 \ --min-p 0.01 \ --top-k 40 \ --ctx-size 16384 \ --seed 3407 ``` {% endcode %} {% endstep %} {% endstepper %} ### 🦙 Llama-server & OpenAI's completion library To deploy MiniMax-M2.5 for production, we use `llama-server` or OpenAI API. In a new terminal say via tmux, deploy the model via: {% code overflow="wrap" %} ```bash ./llama.cpp/llama-server \ --model unsloth/MiniMax-M2.5-GGUF/UD-Q3_K_XL/MiniMax-M2.5-UD-Q3_K_XL-00001-of-00004.gguf \ --alias "unsloth/MiniMax-M2.5" \ --prio 3 \ --temp 1.0 \ --top-p 0.95 \ --min-p 0.01 \ --top-k 40 \ --ctx-size 16384 \ --port 8001 ``` {% endcode %} Then in a new terminal, after doing `pip install openai`, do: {% code overflow="wrap" %} ```python from openai import OpenAI import json openai_client = OpenAI( base_url = "http://127.0.0.1:8001/v1", api_key = "sk-no-key-required", ) completion = openai_client.chat.completions.create( model = "unsloth/MiniMax-M2.5", messages = [{"role": "user", "content": "Create a Snake game."},], ) print(completion.choices[0].message.content) ``` {% endcode %} ## 📊 Benchmarks ### Unsloth GGUF Benchmarks
[Benjamin Marie (third-party) benchmarked](https://x.com/bnjmn_marie/status/2027043753484021810/photo/1) **MiniMax-M2.5** using **Unsloth GGUF quantizations** on a **750-prompt mixed suite** (LiveCodeBench v6, MMLU Pro, GPQA, Math500), reporting both **overall accuracy** and **relative error increase** (how much more often the quantized model makes mistakes vs. the original). Unsloth quants, no matter their precision perform much better than their non-Unsloth counterparts for both accuracy and relative error (despite being 8GB smaller). **Key results:** * **Best quality/size tradeoff here: `unsloth UD-Q4_K_XL`.**\ It’s the closest to Original: only **6.0 points** down, and “only” **+22.8%** more errors than baseline. * **Other Unsloth Q4 quants perform closely together (\~64.5–64.9 accuracy).**\ `IQ4_NL`, `MXFP4_MOE`, and `UD-IQ2_XXS` are all basically the same quality on this benchmark, with **\~33–35%** more errors than Original. * Unsloth GGUFs perform much better than other non-Unsloth GGUFs, e.g. see `lmstudio-community - Q4_K_M` (despite being 8GB smaller) and `AesSedai - IQ3_S`. ### Official Benchmarks You can view further below for benchmarks in table format:
BenchmarkMiniMax-M2.5MiniMax-M2.1Claude Opus 4.5Claude Opus 4.6Gemini 3 ProGPT-5.2 (thinking)
AIME2586.383.091.095.696.098.0
GPQA-D85.283.087.090.091.090.0
SciCode44.441.050.052.056.052.0
IFBench70.070.058.053.070.075.0
AA-LCR69.562.074.071.071.073.0
SWE-Bench Verified80.274.080.980.878.080.0
SWE-Bench Pro55.449.756.955.454.155.6
Terminal Bench 251.747.953.455.154.054.0
HLE w/o tools19.422.228.430.737.231.4
Multi-SWE-Bench51.347.250.050.342.7—
SWE-Bench Multilingual74.171.977.577.865.072.0
VIBE-Pro (AVG)54.242.455.255.636.9—
BrowseComp (w/ctx)76.362.067.884.059.265.8
Wide Search70.363.276.279.457.0—
RISE50.234.050.562.536.850.0
BFCL multi-turn76.837.468.063.361.0—
τ² Telecom97.887.098.299.398.098.7
MEWC74.455.682.189.878.741.3
GDPval-MM59.024.661.173.528.154.5
Finance Modeling21.617.330.133.215.020.0

Coding Core Benchmark Scores

Search and Tool Use

Tasks Completed per 100

Office Capabilities