Is minimax-m2.7-medium good at Data & Analytics?
minimax-m2.7-medium ranks #43 of 69 for Data & Analytics — excellent. The top pick for this task is claude-opus-4.8-low.
minimax-m2.7-medium on each Data & Analytics sub-task
| Spot the Misleading Stat | 99.0/100 | #42 |
| Metric Calculation | 97.0/100 | #31 |
| SQL Reasoning | 91.2/100 | #62 |
| Honest Communication | 86.8/100 | #47 |
Real examples, graded
WeakSimpson's paradox in fill rates (Tradewinds) 11/100
“The model severely failed by fabricating arithmetic and misunderstanding Simpson's paradox, leading to flawed business logic.”
WeakPeeking / optional stopping (Cedar & Sage) 96/100
“The model correctly identifies the core issue as peeking/optional stopping, accurately explains how it inflates the false positive rate, and prescribes the correct methodological solutions (fixed sample size or sequential testing). There is a minor technical misstatement regarding Type II errors in the context of a test that already rejected the null, but the primary reasoning and final conclusion are spot on.”
WeakStatistical vs practical significance (Cedar & Sage) 36/100
“The model fabricated a confidence interval that contradicts the provided p-value and failed to properly evaluate the sample size.”
Frequently asked
Is minimax-m2.7-medium good at Data & Analytics?
minimax-m2.7-medium ranks #43 of 69 models we tested for Data & Analytics, scoring excellent.
What is minimax-m2.7-medium's strongest Data & Analytics skill?
Its best sub-task here is Spot the Misleading Stat.
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.
Prompt × model results
12 test cases · 3 evals