[whatmodel.sh](../../README.md) / MiniMax-M2.5
# MiniMax-M2.5
MiniMax · released 2026-02 · 230B params (10B active) · modified-mit
**score 65.0** ±1.1 · #142 of 732 · from 11 benchmark results across 9 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks MiniMax-M2.5 among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆█▄▅●▄▄▂▁ #108 of 313 20% 3
coding ▁▁▁▃▅▇█▇●▄▃▁ #105 of 469 33% 3
chat & writing ▁▁▁▃▆▄▆█▆●▇▃ #126 of 402* 14% 1
agent tasks · #108 of 313 · 20% of its score
Beats [Grok 4.20 (0309 reasoning)](grok-4-20.md), [o3-pro](o3-pro.md), [Grok 4.3](grok-4-3.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Nova 2 Pro (Preview)](nova-2-pro.md), [Mistral Medium 3.5 (128B)](mistral-medium-3-5.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
ContractBench success rate (33 ob… 62.6% ▆··▆▄▂▄▂▆●█▄ beats 17/23 very hard · avg 13% 11% sharp · at its level
SWE Atlas (Codebase Q&A, Test Wri… 15.2% ●·▄··█···▄·█ beats 0/6 very hard · avg 6.6% 4.6% typical
Vibe Code Bench (test split, 50 w… 14.8% ▃··▃·▃●█▃▃▃█ beats 4/14 very hard · avg 2.8% 4.0% sharp · noisy
coding · #105 of 469 · 33% of its score
Beats [Grok 4.20 (0309 reasoning)](grok-4-20.md), [Grok 4.3](grok-4-3.md), [Muse Glimmer 30B](muse-glimmer-30b.md) here, though they rank higher overall. Trails [Llama 3.1 Nemotron Ultra 253B v1](llama-3-1-nemotron-ultra-253b.md), [MiniMax-M2.1](minimax-m2-1.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena WebDev (Code Arena) 1386.2 ▁▁▃▂▅●▃▆▅▅▁▂ beats 40/129 avg 981 24% sharp
SWE Atlas (Codebase Q&A, Test Wri… 15.2% ●·▄··█···▄·█ beats 0/6 very hard · avg 6.6% 4.6% typical
Vibe Code Bench (test split, 50 w… 14.8% ▃··▃·▃●█▃▃▃█ beats 4/14 very hard · avg 2.8% 4.0% sharp · noisy
chat & writing · #126 of 402 · 14% of its score
Beats [LongCat 2.0 (1.6T-A48B)](longcat-2-0.md), [GPT-5 mini](gpt-5-mini.md) here, though they rank higher overall. Trails [LFM2.5 2.6B](lfm2-5-2-6b.md), [Nemotron-Cascade-8B](nemotron-cascade-8b.md), [DeepSeek V3 0324](deepseek-v3-0324.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LMArena (text) 1390.6 ▁▁▁▂▃▃▃▆▆●█▃ beats 233/385 avg 1260 14% blunt
other tests (6)
test result vs the others in its run difficulty counts why
ARC-AGI-1 public evaluation set (… 59.1% ▁·▇▃▁·●▇▁·▇▃ beats 12/27 very hard · avg 4.0% 10% sharp · at its level · noisy
DSEffi-Bench Pass (1,000 data-sci… 42.8% ▅··▃▅·●·▃██▃ beats 6/15 hard · avg 23% 10% at its level
KINA (Knowledge Index of Noah's A… 30.3% ▁▁▂▁▃▄█●▄▅▂▃ beats 23/39 very hard · avg 21% 9.5% at its level
WildClawBench 27.1% ▃·●▅█·▅█▅▃▃▃ beats 2/17 very hard · avg 11% 6.1% at its level
ARC-AGI-2 public evaluation set (… 5.4% █··▂●▁▃▁▃·▁▃ beats 11/24 very hard · avg 0.3% 5.4% sharp · noisy
Vectara hallucination leaderboard… 90.9% ▁▂▁▂▅█●▃▄▃▁▁ beats 57/98 very easy · avg 90% 0.0% blunt
source results share of the score
LMArena leaderboard dataset 2 38% ▇▇▇▇▇
ARC Prize Foundation model_baseline out… 2 16% ▇▇
ContractBench: Can LLM Agents Preserve … 1 11% ▇
DSEffi-Bench: Demystifying Large Langua… 1 10% ▇
Knowledge Index of Noah's Ark 1 9% ▇
WildClawBench: A Benchmark for Real-Wor… 1 6% ▇
SWE Atlas: Benchmarking Coding Agents B… 1 5% ▇
Vibe Code Bench 1 4% ▇
Vectara hallucination leaderboard 1 0% ▇
prior: pull toward the middle 1% ▇
weighted average = 65.0
Licences: LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · ARC Prize Foundation model\_baseline outputs on ARC-AGI-1/2 public evaluation sets ([MIT](../../licenses/arcprize-model-baseline-outputs-MIT.txt)) · ContractBench: Can LLM Agents Preserve Observation Contracts? ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · DSEffi-Bench: Demystifying Large Language Models' Capability in Efficient Data Science Code Generation ([CC BY 4.0](../../licenses/paper-2608.30248v1-CC-BY-4.0.txt)) · Knowledge Index of Noah's Ark ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · WildClawBench: A Benchmark for Real-World ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · SWE Atlas: Benchmarking Coding Agents Beyond Issue Resolution ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Vibe Code Bench ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### Through an API
api (MiniMax) · $0.52/M blended · $0.3 in / $1.2 out per 1M tokens · [platform.minimax.io](https://platform.minimax.io/docs/guides/text-generation)
[price source](https://platform.minimax.io/docs/guides/pricing-paygo.md): USD per 1M tokens, MiniMax Open Platform pay-as-you-go, MiniMax-M2.5 (listed under 'Legacy Models'): input $0.3, output $1.2. Prompt-cache read $0.03 and write $0.375, and MiniMax-M2.5-highspeed ($0.6/$2.4) excluded. Currency: USD. MiniMax pay-as-you-go pricing page, observed 2026-09-25. · [context source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/models/minimax/MiniMax-M2.5.toml)
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run MiniMax-M2.5 (ud-q4_k_xl, needs ~138.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/MiniMax-M2.5-GGUF/tree/main))
```sh
llama cli -hf unsloth/MiniMax-M2.5-GGUF:UD-Q4_K_XL
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf unsloth/MiniMax-M2.5-GGUF:UD-Q4_K_XL
```
Other versions: ud-q2_k_xl (~69.5 GB) `llama cli -hf unsloth/MiniMax-M2.5-GGUF:UD-Q2_K_XL`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run MiniMax-M2.5 (ud-q4_k_xl, needs ~138.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/MiniMax-M2.5-GGUF/tree/main))
```sh
lms get https://huggingface.co/unsloth/MiniMax-M2.5-GGUF@UD-Q4_K_XL -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: ud-q2_k_xl (~69.5 GB) `lms get https://huggingface.co/unsloth/MiniMax-M2.5-GGUF@UD-Q2_K_XL -y` · 4bit (~128.7 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/MiniMax-M2.5-MLX-4bit -y` · 8bit (~243 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/MiniMax-M2.5-MLX-8bit -y`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run MiniMax-M2.5 (4bit, needs ~128.7 GB, Mac only, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/MiniMax-M2.5-MLX-4bit/tree/9af822e97aaa5d3d8f9a635ec2b967b0c0a9a09f))
```sh
mlx_lm.chat --model lmstudio-community/MiniMax-M2.5-MLX-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model lmstudio-community/MiniMax-M2.5-MLX-4bit
```
Other versions: 8bit (~243 GB, Mac only) `mlx_lm.chat --model lmstudio-community/MiniMax-M2.5-MLX-8bit`
Model repository and documentation: [huggingface.co/MiniMaxAI](https://huggingface.co/MiniMaxAI/MiniMax-M2.5)
[interactive version](https://whatmodel.sh/model/minimax-m2-5) · [JSON](https://whatmodel.sh/api?format=json&all=1)