Grok 4.20 Beta (0309) (Non-reasoning): benchmark results
Provider: xAI. Released 2026-03-09. Access: API.
Unified ELO 1496 ± 1, rank #1490 of 3078 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Natural Intelligence | 39.77 | NatInt Score | 86.3 |
| UGI - Writing | 43.88 | Writing Score | 83.8 |
| UGI Leaderboard | 30.25 | UGI Score | 31.7 |
| UGI - Willingness (W/10) | 4 | W/10 Score | 30.6 |
| ForecastBench Baseline - Zero-Shot - Dataset | 54.3 | Dataset Brier Index (difficulty-adjusted, higher is better) | 11.4 |
| ForecastBench Baseline - Zero-Shot - Market | 52.4 | Market Brier Index (difficulty-adjusted, higher is better) | 5.7 |
| ForecastBench Baseline - Zero-Shot - Overall | 53.3 | Overall Brier Index (difficulty-adjusted, higher is better) | 5.1 |
| ForecastBench | 54.9 | Overall Score (higher is better) | 0.9 |
| ForecastBench Tournament - Market | 55.4 | Market Brier Index (difficulty-adjusted, higher is better) | 0.5 |
Interactive version: theaggregate.ai/model?slug=grok-4-20-beta-0309-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.