Grok 4.5: benchmark results
xAI's flagship Grok 4.5 model for reasoning and general tasks. Provider: xAI. Released 2026-07-08. Access: API.
Aggregate Rating 1824 ± 5, rank #48 of 1656 rated models, from 704 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DataSpace | 66.34 | Task accuracy (%; share of the 410 DataSpace tasks, each a q | 100 |
| E-commerce Last Exam - E-commerce | 78.65 | Mean Reward (%) | 100 |
| EMRB - L5 Spectrum Survey and System Design | 39.8 | Accuracy (%) | 100 |
| HiEviDR-Bench (Deep Research) | 41.18 | Overall score (0-100): sum of five 0-20 dimension scores (mu | 100 |
| HiEviDR-Bench - Arxiv (Deep Research) | 35.95 | Overall score (0-100): sum of five 0-20 dimension scores (mu | 100 |
| HiEviDR-Bench - Wiki (Deep Research) | 46.41 | Overall score (0-100): sum of five 0-20 dimension scores (mu | 100 |
| Hy-MultiTurn - Constraint Synthesis | 90 | Importance-weighted score (%) on the 35 constraint synthesis | 100 |
| Hy-MultiTurn - Object Localization | 95.2 | Importance-weighted score (%) on the 35 object localization | 100 |
| Hy-MultiTurn - Reference Resolution | 78 | Importance-weighted score (%) on the 35 reference resolution | 100 |
| LLM Stats (AutomationBench-AA) | 51.4 | Score (%) | 100 |
| Long-Horizon Terminal-Bench | 50.5 | Mean Score (%) | 100 |
| Nejumi 4 - ALT - Controllability | 94.73 | Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=grok-4-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.