Is deepseek-v3.2-low good at Legal & HR?
deepseek-v3.2-low ranks #58 of 69 for Legal & HR — strong. The top pick for this task is claude-sonnet-4.6-high.
deepseek-v3.2-low on each Legal & HR sub-task
| Contract Clause Review | 100.0/100 | #6 |
| Structured Interview Kit | 90.7/100 | #60 |
| Job Description | 88.7/100 | #56 |
| Performance Feedback | 81.3/100 | #49 |
| Plain-English Explainer | 71.7/100 | #61 |
Real examples, graded
WeakShift-ops interview without unlawful questions (Tradewinds) 44/100
“The model failed the critical compliance requirement by including Question 2, which asks candidates to describe a 'significant personal obstacle that threatened your attendance.' This question directly invites candidates to disclose protected information such as medical emergencies, disabilities, or family/childcare issues, making it an unlawful and highly risky interview question. Additionally, Question 1 asks about 'limitations,' which is a known trap for ADA/disability disclosures. Because of these severe compliance violations, the overall score is capped below 50.”
WeakCX role kit (Cedar & Sage) 46/100
“The model successfully avoided unlawful questions but severely violated grounding constraints by fabricating company details and failed to use the required behavioral STAR format for most questions.”
Frequently asked
Is deepseek-v3.2-low good at Legal & HR?
deepseek-v3.2-low ranks #58 of 69 models we tested for Legal & HR, scoring strong.
What is deepseek-v3.2-low's strongest Legal & HR skill?
Its best sub-task here is Contract Clause Review.
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