xAIGrok
Grok 4.20 beta reasoning
xAI · Grok. Released Mar 9, 2026.
Grok 4.20 beta reasoning is a model from xAI in the Grok family, released Mar 9, 2026. evals.report tracks 12 reported Grok 4.20 beta reasoning benchmark scores across ARC-AGI-1, ARC-AGI-3, LMArena, LiveBench, ARC-AGI-2, WeirdML, Epoch Capabilities Index, Global-MMLU, and 4 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
12 results
Benchmark results 12
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| ARC-AGI-1 | Reasoning | 89.5% | accuracy | Official | Mar 9, 2026 | Details |
| ARC-AGI-3 | Reasoning | 0.09% | accuracy | Official | Mar 9, 2026 | Details |
| LMArena | Chat preference | 1453 | source-defined rating | Official | Mar 9, 2026 | Details |
| LiveBench | Reasoning | 67.96% | score | Official | Mar 9, 2026 | Details |
| ARC-AGI-2 | Reasoning | 65.14% | accuracy | Official | Mar 9, 2026 | Details |
| WeirdML | Coding | 52.3% | average accuracy | Official | Mar 9, 2026 | Details |
| Epoch Capabilities Index | Reasoning | 154.3 | Index | Official | Mar 9, 2026 | Details |
| Global-MMLU | Reasoning | 89.5% | accuracy | Unverified | Mar 9, 2026 | Details |
| WebDev Arena | Chat preference | 1395 | Elo | Verified | Mar 9, 2026 | Details |
| Search Arena | Chat preference | 1193 | Elo | Verified | Mar 9, 2026 | Details |
| EQ-Bench Creative Writing v3 | Chat preference | 1675 | Elo | Verified | Mar 9, 2026 | Details |
| Design Arena | Chat preference | 1271 | Elo | Verified | Mar 9, 2026 | Details |