Ex0bit / MiniMax-M2.5-PRISM-LITE

MiniMax-M2.5-PRISM-LITE

A PRISM-LITE version of Ex0bit/MiniMax-M2.5-PRISM-PRO intended for role-following over-refusal and propaganda mechanisms suppression using our SOTA PRISM pipeline.

PRISM-PRO version available for purchase here: https://ko-fi.com/s/0a23d1b9a5

For Full Custom trained PRISM versions or raw tensors access reach out @ https://ko-fi.com/ex0bit .

Model Highlights

• PRISM Ablation — State-of-the-art technique that removes over-refusal behaviors while preserving model capabilities
• SOTA Coding Performance — 80.2% on SWE-Bench Verified, 51.3% on Multi-SWE-Bench, 76.3% on BrowseComp (with context management)
• Frontier Agentic Capabilities — Industry-leading performance in tool use, search, and complex multi-step tasks
• Efficient Reasoning — Trained with RL to reason efficiently and decompose tasks optimally, 37% faster than M2.1
• Cost-Effective — $1 for continuous operation at 100 tok/s for an hour; $0.30 at 50 tok/s
• Modified-MIT Base License — Based on MiniMax's open-weight release

Base Model Architecture

MiniMax-M2.5 is a Mixture-of-Experts (MoE) model extensively trained with reinforcement learning across hundreds of thousands of complex real-world environments.

Benchmarks (Base Model)

Coding

Search & Tool Calling

Reasoning & Knowledge

Usage

llama.cpp (GGUF)

Build the latest master of llama.cpp and run:

~/llama.cpp/build/bin/llama-cli \
  -m ../outputs/MiniMax-M2.5-PRISM-PRO-[QUANT].gguf \
  --jinja \
  -ngl 999 \
  --repeat_penalty 1.15 \
  --temp 1.0 \
  --top_p 0.95 \
  --top_k 40

Replace [QUANT] with your quantization level (e.g. Q8_0 , etc.).

Recommended Parameters

License

This model is released under the PRISM Research License .

The base model MiniMax-M2.5 is released under a Modified-MIT License .

Acknowledgments

Based on MiniMax-M2.5 by MiniMax AI .