Gemini 4 Argon, made generally available in mid-October 2026, now anchors trader views as Google’s strongest 2026 performer on Humanity’s Last Exam variants, posting 57.1% on Artificial Analysis HLE and 56.9% on the HLE-Diamond subset without tools. These results reflect targeted gains in complex reasoning, coding, and scientific tasks, yet trail Anthropic’s Claude Fable 5.1 and Opus 5.5 models that lead most public leaderboards at 59–65%. Earlier Gemini 3.8 Flash entries reached the low-to-mid 50s on verified subsets, underscoring rapid iteration within Google’s lineup. Key catalysts ahead include full consumer rollout, additional agentic tooling updates, and any new model variants before year-end that could lift the highest Gemini score.