Alibaba Qwen 2.5 72B vs MiniMax MiniMax M2

Detailed comparison for LLMs

AlibabaMiniMax

On Guardion's LLM vulnerability Benchmark, Alibaba Qwen 2.5 72B is the more secure of the two: Qwen 2.5 72B scores 26.3% and MiniMax M2 scores 40.0% on attack success rate (ASR) (lower is better). One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.

Head-to-Head Overview

Qwen 2.5 72B is the overall winner in this comparison!

Attack Success Rate (lower is safer)

ASR for Alibaba Qwen 2.5 72B vs MiniMax MiniMax M2. Green marks the safer model on each metric. Only the overall score is available for estimated models.

Overall (ASR)

Qwen 2.5 72B
26.3%
MiniMax M2
40.0%

TAP Attack Method (ASR)

Qwen 2.5 72B
55.2%
MiniMax M2
100.0%

Crescendo Attack Method (ASR)

Qwen 2.5 72B
23.7%
MiniMax M2
100.0%

Zero-Shot (ASR)

Qwen 2.5 72B
0.3%
MiniMax M2
100.0%

Key Highlights

  • Alibaba Qwen 2.5 72B has a lower Overall (ASR).
  • Alibaba Qwen 2.5 72B has a lower TAP Attack Method (ASR).
  • Alibaba Qwen 2.5 72B has a lower Crescendo Attack Method (ASR).
  • Alibaba Qwen 2.5 72B has a lower Zero-Shot (ASR).

Security Profile

Outward is better on every axis.

OverallTAPCrescendoZero-Shot
Qwen 2.5 72B
MiniMax M2
Full security profile
Alibaba Qwen 2.5 72B
Full security profile
MiniMax MiniMax M2

Frequently asked questions

Is Alibaba Qwen 2.5 72B or MiniMax MiniMax M2 more secure?

On Guardion's LLM vulnerability Benchmark, Alibaba Qwen 2.5 72B is the more secure of the two: Qwen 2.5 72B scores 26.3% and MiniMax M2 scores 40.0% on attack success rate (ASR) (lower is better). One or both scores are estimated from public safety evaluations pending a Guardion benchmark run.

What is the attack success rate (ASR) of Qwen 2.5 72B vs MiniMax M2?

Qwen 2.5 72B has a 26.3% ASR and MiniMax M2 has a 40.0% ASR — the share of adversarial prompts that succeed across zero-shot, TAP, and Crescendo attacks. Lower is safer.

How were Qwen 2.5 72B and MiniMax M2 tested?

Both were red-teamed with the HarmBench framework across zero-shot, TAP (Tree of Attacks with Pruning), and Crescendo multi-turn attacks, scored by Attack Success Rate.

Related Comparisons