Rankings / xAI
Grok 4.20 0309 Reasoning
released 2026-02-17
BenchAtlas Index
as of 2026-09-11
base configuration
rank #120
8 families · 3 categories · medium
Benchmark evidence
15 results
external indices
| ECI | 152.0 points | independent | Epoch AI Benchmarking Hub | |
| Vals Index | 17.6 points | independent | Vals AI |
factuality
| SimpleQA Verified | 35.1 % | independent | Epoch AI Benchmarking Hub 2026-07-13 | |
| SimpleQA Verified | 30.2 % | independent | Epoch AI Benchmarking Hub 2026-08-27 |
knowledge science
| GPQA Diamond | 89.3 % | independent | Epoch AI Benchmarking Hub 2026-07-13 | |
| GPQA Diamond implementation=vals-ai | 88.6 % | independent | Vals AI | |
| MMLU-Pro implementation=vals-ai | 86.3 % | independent | Vals AI |
multimodal
| MMMU implementation=vals-ai | 83.5 % | independent | Vals AI |
professional
| CaseLaw | 54.4 % | independent | Vals AI | |
| CorpFin | 63.7 % | independent | Vals AI | |
| LegalBench | 77.7 % | independent | Vals AI | |
| MedQA | 94.5 % | independent | Vals AI | |
| TaxEval | 74.1 % | independent | Vals AI |
reasoning math
| AIME implementation=vals-ai | 96.5 % | independent | Vals AI | |
| OTIS Mock AIME 2024–2025 | 92.2 % | independent | Epoch AI Benchmarking Hub 2026-07-13 |
Agent + model results
systems, not bare-model scores
| agent + model grok-cli + Grok 4.20 0309 Reasoning | Terminal-Bench 2.0 | 57.3 % | unverified | Terminal-Bench Leaderboard |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.