DeepSeek V4 Pro (0424): benchmark results
Provider: DeepSeek. Released 2026-04-24. Access: Open.
Aggregate Rating 1786 ± 32, rank #116 of 10916 rated models, from 56 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA TAU-2 Bench | 96.2 | Accuracy (%) | 96.8 |
| AA IFBench | 76.46 | Accuracy (%) | 95.6 |
| EnterpriseOps-Gym-AA - Hybrid | 43.14 | Task Success (%) | 93 |
| AA Terminal-Bench Hard | 46.21 | Accuracy (%) | 92.9 |
| AA GPQA Diamond | 90.51 | Accuracy (%) | 90.3 |
| AutomationBench-AA - Finance - Raw Objective Completion | 86.56 | Raw Objective Completion (%) | 84.6 |
| AA Humanity's Last Exam | 37.53 | Accuracy (%) | 83.6 |
| AA CritPt | 12.86 | Accuracy (%) | 80.3 |
| ReLE - Coding | 72.2 | Accuracy (%) | 76.5 |
| AutomationBench-AA - Strict | 26.18 | Strict Task Success (%) | 76.1 |
| AutomationBench-AA | 56.29 | Guardrail-Adjusted Objective Completion (%) | 73.7 |
| AutomationBench-AA - Operations - Raw Objective Completion | 91.77 | Raw Objective Completion (%) | 72.5 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-pro-0424 · How It Works · Data refreshed daily, snapshot 2026-11-08.