DeepSeek V4 Pro
Weighted composite
4.48
Recommendation
Practitioner-grade
Cohort
large open weight
Scorecard
Per-criterion scores not available for this model. It appears in the cross-part overview but is not part of the canonical Models 1–10 hand-graded set; only the headline composite is shown.
Per-part composites
| Part | Opus 4.7 (inline) | DeepSeek V4 Pro (judge) |
|---|---|---|
| Part A | 4.48 | 4.00 |
| Part B | 4.50 | 3.55 |
| Part C | 4.65 | 4.45 |
Notes from the evaluation
DeepSeek V4 Pro second-judge outputs from `runner/grade.py`. - Part B Opus scores are sampled estimates unless the file explicitly says a full per-question pass was performed. - Part B DeepSeek scores are the current whole-response judge scores.
Source files in the repo
Cross-judge composites: analysis/results_overview.md
Full report · PDF
Get the full report
All 111 questions, the complete rubric, per-model verdicts, and the methodology paper. Delivered as PDF within 60 seconds. CC-BY 4.0.
Welcome back.
You've already requested the full report.