Confirm Action

Are you sure you want to proceed?

Is Minimax m2.7 good at Customer Support?

Minimax m2.7 ranks #35 of 43 for Customer Support — needs editing. The top pick for this task is gemini-3.1-pro-preview-low.

#35 / 43
Rank for this task
62.4
Score
$0.0202
Cost / run

Minimax m2.7 on each Customer Support sub-task

Escalation and Incident Test 70.0/100 #34
Escalation & Handoff 69.5/100 #24
Help Content Test 63.6/100 #43
Basic Support Reply Test 60.4/100 #33
De-escalation 60.0/100 #29
Policy Boundaries 59.7/100 #43
Resolution 53.5/100 #35
Policy Boundary Test 53.5/100 #35

Real examples, graded

WeakDuplicate charge (Cedar & Sage) 52/100

“The model correctly identifies the account facts and policy timeline, but makes a severe mathematical error by promising a $68 refund. Since the customer was charged $34 twice, only one $34 charge is a duplicate to be reversed. Promising $68 is an unauthorized financial commitment.”

WeakPassword reset help 45/100

“The model failed to write a ready-to-use customer support reply. Instead, it generated a generic knowledge base article with placeholders and referred to the company in the third person. It completely lacks the empathy, greeting, and conversational tone expected in a support interaction.”

WeakAbuse boundary 52/100

“The model provides the correct workarounds but fails significantly by inventing a commitment to prioritize a fix. Furthermore, it literally translates the internal escalation policy into a customer-facing threat, which is poor support practice and likely to escalate hostility.”

← Full Minimax m2.7 review All Customer Support rankings → Top pick: gemini-3.1-pro-preview-low →

Frequently asked

Is Minimax m2.7 good at Customer Support?

Minimax m2.7 ranks #35 of 43 models we tested for Customer Support, scoring needs editing.

What is Minimax m2.7's strongest Customer Support skill?

Its best sub-task here is Escalation and Incident Test.

This page is Spring Prompt, running

We just did this for every model. Do it for your prompt.

The rankings above come from running real tasks through real models and scoring every output. Spring Prompt is that same engine — pointed at your prompt, your test cases, and your definition of good.

  • Generate test cases from your prompt — no eval set required to start.
  • Compare models side by side with quality, cost and latency in one matrix.
  • Optimise the winner until the scores say it's ready to ship.
Experiment · Cold outreach email

Prompt × model results

12 test cases · 3 evals
Claude Opus
GPT-5
Gemini
v1
7.1
6.8
7.4
v2
8.3
7.9
8.0
v3
9.2
8.6
8.4
Best combo: v3 × Claude Opus
9.2 quality · $0.004/run · 1.8s