Grok 4.5 — benchmark results
Provider: xAI. Access: API.
Unified ELO 1923 ± 26, rank #41 of 1719 rated models, from 30 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SWE-Marathon | 29 | Pass@1 (%) | 100 |
| Vals AI Harvey Legal Agent Bench | 12.92 | Accuracy (%) | 100 |
| UGI Leaderboard | 62.32 | UGI Score | 99.1 |
| Vals AI GPQA | 92.93 | Accuracy (%) | 97.4 |
| Vals AI SWE-bench Verified | 86.6 | Resolved (%) | 96.9 |
| Vals AI MedScribe | 86.88 | Accuracy (%) | 95.5 |
| UGI - Natural Intelligence | 60.21 | NatInt Score | 95.4 |
| Vals AI LiveCodeBench | 87.35 | Accuracy (%) | 95.1 |
| AI Chess Leaderboard (Reasoning) | 1516 | Elo | 94.7 |
| UGI - Writing | 59.96 | Writing Score | 93.9 |
| SvelteBench | 98.9 | Average pass@1 (%) | 93.5 |
| VoxelBench | 1796 | Rating | 93 |
Interactive version: aibenchmarks.dev/model?slug=grok-4-5 · How the rankings work · Data refreshed daily, snapshot 2026-07-09.