Grok 4.7 (Medium): benchmark results
Provider: xAI. Released 2026-09-21. Access: API.
Aggregate Rating 1866 ± 21, rank #56 of 2163 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DuelLab Overall | 73.24 | GameBench2 normalized score (0–100) | 88.1 |
| Argo-Bench | 36.6 | Mean task score (0-100) over 210 enterprise data-science tas | 76.7 |
| Argo-Bench - Solved | 16.67 | Share of the 210 tasks solved (%), a task counting as solved | 74.2 |
| ARC-AGI-1 | 90 | Accuracy (%) | 72.2 |
| ARC-AGI-2 | 58.75 | Accuracy (%) | 67.6 |
| ARC-AGI-3 | 1.71 | Accuracy (%) | 61.2 |
| NonoBench | 76.67 | Standard accuracy (%), correct / non-failed attempts; versio | 61 |
| Bug Hunt Bench | 26.7 | Planted Bugs Fixed (out of 105) | 58.4 |
| Bug Hunt Bench - LMS | 14.7 | Planted Bugs Fixed (out of 60) | 57.6 |
| Bug Hunt Bench - VS Code Extension | 12 | Planted Bugs Fixed (out of 45) | 57.6 |
| CursorBench 4.0 | 41.6 | Score (%) | 53 |
| MedSafe-Dx v0.3 (4aj) | 60.8 | Balanced escalation utility (fixed-flag policy = 0, perfect | 52.6 |
Interactive version: theaggregate.ai/model?slug=grok-4-7-medium · How It Works · Data refreshed daily, snapshot 2026-10-24.