X @Elon Musk

And the latest Grok 4.20 checkpoints are much better. Largest model variant of 4.20 still hasn’t finished training.X Freeze (@XFreeze):Grok 4.20 (Preview) from xAI just ranked #2 on ForecastBench’s global AI forecasting leaderboardOutperforming GPT‑5, Gemini 3 Pro, and Claude Opus 4.5, while closing in on elite human superforecasters https://t.co/DiezD6rWGw ...