Deepseek Deepseek R1 vs MiniMax MiniMax M1 80K

Detailed comparison for LLMs

DeepseekMiniMax

On Guardion's LLM vulnerability Benchmark, Deepseek Deepseek R1 is the more secure of the two: Deepseek R1 scores 42.6% and MiniMax M1 80K scores 43.0% 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

Deepseek R1 is the overall winner in this comparison!

Attack Success Rate (lower is safer)

ASR for Deepseek Deepseek R1 vs MiniMax MiniMax M1 80K. Green marks the safer model on each metric. Only the overall score is available for estimated models.

Overall (ASR)

Deepseek R1
42.6%
MiniMax M1 80K
43.0%

TAP Attack Method (ASR)

Deepseek R1
76.9%
MiniMax M1 80K
100.0%

Crescendo Attack Method (ASR)

Deepseek R1
34.8%
MiniMax M1 80K
100.0%

Zero-Shot (ASR)

Deepseek R1
16.1%
MiniMax M1 80K
100.0%

Key Highlights

  • Deepseek Deepseek R1 has a lower Overall (ASR).
  • Deepseek Deepseek R1 has a lower TAP Attack Method (ASR).
  • Deepseek Deepseek R1 has a lower Crescendo Attack Method (ASR).
  • Deepseek Deepseek R1 has a lower Zero-Shot (ASR).

Security Profile

Outward is better on every axis.

OverallTAPCrescendoZero-Shot
Deepseek R1
MiniMax M1 80K
Full security profile
Deepseek Deepseek R1
Full security profile
MiniMax MiniMax M1 80K

Frequently asked questions

Is Deepseek Deepseek R1 or MiniMax MiniMax M1 80K more secure?

On Guardion's LLM vulnerability Benchmark, Deepseek Deepseek R1 is the more secure of the two: Deepseek R1 scores 42.6% and MiniMax M1 80K scores 43.0% 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 Deepseek R1 vs MiniMax M1 80K?

Deepseek R1 has a 42.6% ASR and MiniMax M1 80K has a 43.0% ASR — the share of adversarial prompts that succeed across zero-shot, TAP, and Crescendo attacks. Lower is safer.

How were Deepseek R1 and MiniMax M1 80K 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