Google announced Gemini 4 Carbon on October 28, 2026, positioning the large language model as its strongest frontier release yet with gains in long-horizon reasoning, secure coding, and cybersecurity benchmarks. Independent evaluator Artificial Analysis immediately ranked it first on the new HLE-Diamond leaderboard at 69.1 percent overall—79.6 percent on reasoning and 58.6 percent on knowledge—edging Claude Opus 5.5 and establishing a clear lead in the Intelligence Index v5. The model remains in limited Fairwind Program access for vetted cyber defenders, with no public API yet, mirroring the phased rollout of sibling Gemini 4 Argon. Trader consensus prices high implied probabilities for debuts above 65 percent while assigning only 21 percent odds to clearing 70 percent, reflecting the verified initial score and the competitive landscape where subsequent models have posted higher results. Broader developer availability and any follow-on benchmark updates remain key near-term catalysts.